Recent changes¶
Version 8.16¶
Changes in 8.16.0¶
Kernel¶
Changed: Fixpoints are now expected to be guarded even in subterms erasable by reduction, thus getting rid of an artificial obstacle preventing to lift the assumption of weak normalization of Coq to an assumption of strong normalization; for instance (barring implementation bugs) termination of the type-checking algorithm of Coq is now restored (of course, as usual, up to the assumption of the consistency of set theory and type theory, i.e., equivalently, up to the weak normalization of type theory, a "physical" assumption, which has not been contradicted for decades and which specialists commonly believe to be a truth) (#15434, incidentally fixes the complexity issue #5702, by Hugo Herbelin).
Changed: Flag
Unset Guard Checking
nevertheless requires fixpoints to have an argument marked as decreasing in a type which is inductive (#15668, fixes #15261, by Hugo Herbelin).Removed: Template polymorphism is now forbidden for mutual inductive types (#15965, by Gaëtan Gilbert).
Fixed: Inlining of non-logical objects (notations, hints, ...) was missing when applying a functor returning one of its arguments as e.g. in
Module F (E:T) := E
(#15412, fixes #15403, by Hugo Herbelin).Fixed: We introduce a new irrelevant term in the reduction machine. It is used to shortcut computation of terms living in a strict proposition, and behaves as an exception. This restores subject reduction, and also makes conversion of large terms in SProp cheap (#15575, fixes #14015, by Pierre-Marie Pédrot).
Fixed: performance blowups while inferring variance information for Cumulative, NonCumulative inductive types (#15662, fixes #11741, by Gaëtan Gilbert).
Specification language, type inference¶
Added: New clause
as ident
to theRecord
command to specify the name of the main argument to use by default in the type of projections (#14563, by Hugo Herbelin).Added: Reversible coercions are coercions which cannot be represented by a regular coercion (a Gallina function) but rather a meta procedure, such as type class inference or canonical structure resolution (#15693, by Cyril Cohen, Pierre Roux, Enrico Tassi, reviewed by Ali Caglayan, Jim Fehrle and Gaëtan Gilbert).
Added: support for coercions not fulfilling the uniform inheritance condition, allowing more freedom for the parameters that are now inferred using unification, canonical structures or typeclasses (#15789, fixes #2828, #4593, #3115, #5222, #9696 and #8540, by Pierre Roux, reviewed by Ali Caglayan, Enrico Tassi, Kazuhiko Sakaguchi and Jim Fehrle).
Fixed: interpretation of
{struct}
fixpoint annotations when the principal argument comes from an implicit generalization (#15581, fixes #13157, by Gaëtan Gilbert).
Notations¶
Removed:
_
inident
entries in notations, which was deprecated in favor ofname
in 8.13. When you see messages likeError: Notation "[ rel _ _ : _ | _ ]" is already defined at level 0 with arguments name, name, constr, constr while it is now required to be at level 0 with arguments ident, ident, constr, constr.
replace
ident
withname
in theNotation
command. To ease the change, you can fix thedeprecated-ident-entry
warnings in Coq 8.15 (or 8.14 or 8.13). The warning can be turned into an error with-arg -w -arg +deprecated-ident-entry
in the_CoqProject
file (#15754, by Pierre Roux).Added: When defining a recursive notation referring to another recursive notation, expressions of the form
x .. y
can be used where a sequence of binders is expected (#15291, grants #7911, by Hugo Herbelin).Fixed: Coercions are disabled when typechecking parsers and printers of
Number Notation
(#15884, fixes #15843, by Pierre Roux).
Tactics¶
Changed: The
RewriteRelation
type class is now used to declare relations inferrable by thesetoid_rewrite
tactic to constructProper
instances. This can break developments that relied on existingReflexive
instances to infer relations. The fix is to simply add a (backwards compatible)RewriteRelation
declaration for the relation. This change allows to set stricter modes on the relation type classesReflexive
,Symmetric
, etc. (#13969, fixes #7916, by Matthieu Sozeau).Changed: The
setoid_rewrite
tactic can now properly recognize homogeneous relations applied to types in different universes. (#14138, fixes #13618, by Matthieu Sozeau).Changed: The
eauto
tactic does not propagate internal Ltac failures with level > 0 anymore. Any failure caused by a hint now behaves as if it were a level 0 error (#15215, fixes #15214, by Pierre-Marie Pédrot).Changed:
rewrite
when used to rewrite in multiple hypotheses (egrewrite foo in H,H'
) requires that the term (foo
) does not depend on the hypotheses it rewrites. When usingrewrite in *
, this means we only rewrite in hypotheses which do not appear in the term. (#15426, fixes #3051 and #15448, by Gaëtan Gilbert).Changed: When it fails,
assert_succeeds
fails with the argument tactic's original error instead ofTactic failure: <tactic closure> fails.
(#15728, fixes #10970, by Gaëtan Gilbert).Deprecated: the
instantiate
tactic without arguments. Since the move to the monadic tactic engine in 8.5, it was behaving as the identity (#15277, by Pierre-Marie Pédrot).Added: generalized rewriting now supports rewriting with (possibly polymorphic) relations valued in
Type
. UseClasses.CMorphisms
instead ofClasses.Morphisms
to declareProper
instances forrewrite
(orsetoid_rewrite
) to use when rewriting withType
valued relations. (#14137, fixes #4632, #5384, #5521, #6278, #7675, #8739, #11011, #12240, and #15279, by Matthieu Sozeau helped by Ali Caglayan).Added: Tactics to obtain a micromega cone expression (aka witness) from an already reified goal. Using those tactics, the user can develop their own micromega tactics for their own types, using their own parsers (#15921, by Pierre Roux, reviewed by Frédéric Besson and Jim Fehrle).
Fixed:
typeclasses eauto
used with multiple hint databases respects priority differences for hints from separate databases (#15289, fixes #5304, by Gaëtan Gilbert).Fixed:
cbn
has better support for combiningsimpl nomatch
,!
and/
specifiers (c.f.Arguments
) (#15657, fixes #3989 and #15206, by Gaëtan Gilbert).
Tactic language¶
Changed: Ltac
match
does not fail when the term to match contains an unfolded primitive projection (#15559, fixes #15554, by Gaëtan Gilbert).Added:
Ltac2
understandstoplevel_selector
and obeysDefault Goal Selector
. Note thatpar:
is buggy when combined withabstract
. UnlikeLtac1
evenpar: abstract tac
is not properly treated. (#15378, by Gaëtan Gilbert).Added: Ltac2
Int
functionsdiv
,mod
,asr
,lsl
,lsr
,land
,lor
,lxor
andlnot
. (#15637, by Michael Soegtrop).Fixed: Ltac2
apply
andeapply
not unifying with implicit arguments; unification inconsistent withexact
andeexact
(#15741, by Ramkumar Ramachandra).
Commands and options¶
Changed:
Module
now only allows parentheses around module arguments. For instance,Module M := (F X).
is now a parsing error (#15355, by Gaëtan Gilbert).Changed:
Fail
no longer catches anomalies, which it has done since Coq version 8.11. Now it only catches user errors (#15366, by Hugo Herbelin).Changed: Program Definition in universe monomorphic mode does not accept non-extensible universe declarations (#15424, fixes #15410, by Gaëtan Gilbert).
Changed: The algorithm for name generation of anonymous variables for
Program
subproofs is now the same as the one used in the general case. This can create incompatibilities in scripts relying on such autogenerated names. The old scheme can be reactivated using the deprecated flagProgram Naming
(#15442, by Pierre-Marie Pédrot).Removed:
Universal Lemma Under Conjunction
flag, that was deprecated in 8.15 (#15268, by Théo Zimmermann).Removed:
Abort
no longer takes anident
as an argument (it has been ignored since 8.5) (#15669, by Gaëtan Gilbert).Removed:
Simplex
flag, that was deprecated in 8.14.lia
andlra
will always use the simplex solver (that was already the default behaviour). (#15690, by Frédéric Besson).Deprecated:
Add LoadPath
andAdd Rec LoadPath
. If this command is an important feature for you, please open an issue onGitHub <https://github.com/coq/coq/issues>
and explain your workflow (#15652, by Gaëtan Gilbert).Deprecated: the
Typeclasses Filtered Unification
flag. Due to a buggy implementation, it is unlikely this is used in the wild (#15752, by Pierre-Marie Pédrot).Added:
Scheme Boolean Equality
command to generate the boolean equality for an inductive type whose equality is decidable. It is useful when Coq is able to generate the boolean equality but isn't powerful enough to prove the decidability of equality (unlikeScheme Equality
, which tries to prove the decidability of the type) (#15526, by Hugo Herbelin).Added: New more extensive algorithm based on the "parametricity" translation for canonically generating Boolean equalities associated to a decidable inductive type (#15527, by Hugo Herbelin).
Added:
From … Dependency
command to declare a dependency of a.v
file on an external file. Thecoqdep
tool generates build dependencies accordingly (#15650, fixes #15600, by Enrico Tassi).Added:
Print Notation
command that prints the level and associativity of a given notation definition string (#15683, fixes #14907 and #4436 and #7730, by Ali Caglayan and Ana Borges, with help from Emilio Jesus Gallego Arias).Added: a warning when trying to deprecate a definition (#15760, by Pierre Roux).
Added: A deprecation warning that
>
in a field of aClass
and theClass >
syntax will in the future also declare coercions as they do when used inRecord
commands (#15802, addresses #2643 and #9014, by Pierre Roux, reviewed by Gaëtan Gilbert, Ali Caglayan, Jason Gross, Jim Fehrle and Théo Zimmermann).Added: the
nonuniform
boolean attribute that silences the non-uniform-inheritance warning when user needs to declare such a coercion on purpose (#15853, by Pierre Roux, reviewed by Gaëtan Gilbert and Jim Fehrle).Added: All commands which can import modules (e.g.
Module Import M.
,Module F (Import X : T).
,Require Import M.
, etc) now supportimport_categories
.Require Import
andRequire Export
also supportfiltered_import
. (#15945, fixes #14872, by Gaëtan Gilbert).Fixed: Make
Require Import M.
equivalent toRequire M. Import M.
(#15347, fixes #3556, by Maxime Dénès).
Command-line tools¶
Added: coq_makefile variable
COQPLUGININSTALL
to configure the installation of ML plugins (#15788, by Cyril Cohen and Enrico Tassi).Added: Added
-bytecode-compiler yesno
flag forcoqchk
enablingvm_compute
during checks, which is off by default. (#15886, by Ali Caglayan).Fixed:
coqdoc
confused by the presence of commandLoad
in a file (#15511, fixes #15497, by Hugo Herbelin).
Standard library¶
Changed: the
signature
scope ofClasses.CMorphisms
intosignatureT
(#15446, by Olivier Laurent).Changed: the locality of typeclass instances
Permutation_app'
andPermutation_cons
fromglobal
toexport
(#15597, fixes #15596, by Gaëtan Gilbert).Removed:
Int63
, which was deprecated in favor ofUint63
in 8.14 (#15754, by Pierre Roux).Deprecated: some obsolete files from the
Arith
part of the standard library (Div2
,Even
,Gt
,Le
,Lt
,Max
,Min
,Minus
,Mult
,NPeano
,Plus
). ImportArith_base
instead of these files. References to items in the deprecated files should be replaced with references toPeanoNat.Nat
as suggested by the warning messages. Concerning the definitions of parity properties (even and odd), it is recommended to useNat.Even
andNat.Odd
. If an inductive definition of parity is required, the mutually inductiveNat.Even_alt
andNat.Odd_alt
can be used. However, induction principles forNat.Odd
andNat.Even
are available asNat.Even_Odd_ind
andNat.Odd_Even_ind
. The equivalence between the non-inductive and mutually inductive definitions of parity can be found inNat.Even_alt_Even
andNat.Odd_alt_Odd
. AllHint
declarations in thearith
database have been moved toArith_prebase
andArith_base
. To use the results about Peano arithmetic, we recommend importingPeanoNat
(orArith_base
to base it on thearith
hint database) and using theNat
module.Arith_prebase
has been introduced temporarily to ensure compatibility, but it will be removed at the end of the deprecation phase, e.g. in 8.18. Its use is thus discouraged. (#14736, #15411, by Olivier Laurent, with help of Karl Palmskog).Deprecated:
identity
inductive (replaced by the equivalenteq
).Init.Logic_Type
is removed (the only remaining definitionnotT
is moved toInit.Logic
) (#15256, by Olivier Laurent).Deprecated:
P_Rmin
: use more generalRmin_case
instead (#15388, fixes #15382, by Olivier Laurent).Added: lemma
count_occ_rev
(#15397, by Olivier Laurent).Added:
Nat.EvenT
andNat.OddT
(almost the same asNat.Even
andNat.Odd
but with output inType
. Decidability of parity (with outputType
) is providedEvenT_OddT_dec
as well as induction principlesNat.EvenT_OddT_rect
andNat.OddT_EvenT_rect
(with outputType
) (#15427, by Olivier Laurent).Added: Added a proof of
sin x < x
for positivex
andx < sin x
for negativex
(#15599, by stop-cran).Added: decidability typeclass instances for Z.le, Z.lt, Z.ge and Z.gt, added lemmas Z.geb_ge and Z.gtb_gt (#15620, by Michael Soegtrop).
Added: lemmas
Rinv_inv
,Rinv_mult
,Rinv_opp
,Rinv_div
,Rdiv_opp_r
,Rsqr_div'
,Rsqr_inv'
,sqrt_inv
,Rabs_inv
,pow_inv
,powerRZ_inv'
,powerRZ_neg'
,powerRZ_mult
,cv_infty_cv_0
, which are variants of existing lemmas, but without any hypothesis (#15644, by Guillaume Melquiond).Added: a Leibniz equality test for primitive floats (#15719, by Pierre Roux, reviewed by Guillaume Melquiond).
Added: support for primitive floats in Scheme Boolean Equality (#15719, by Pierre Roux, reviewed by Hugo Herbelin).
Infrastructure and dependencies¶
Changed: Bumped lablgtk3 lower bound to 3.1.2 (#15947, by Pierre-Marie Pédrot).
Changed: Load plugins using findlib. This requires projects built with
coq_makefile
to either provide a hand writtenMETA
file or use the-generate-meta-for-package
option when applicable. As a consequenceDeclare ML Module
now uses plugin names according tofindlib
, e.g.coq-aac-tactics.plugin
.coqdep
accepts-m META
and uses the file to resolve plugin names to actual file names. (#15220, fixes #7698, by Enrico Tassi).Changed: Minimum supported zarith version is now 1.11 (#15483 and #16005 and #16030, closes #15496, by Gaëtan Gilbert and Théo Zimmermann and Jason Gross).
Changed: Bump the minimum OCaml version to 4.09.0. As a consequence the minimum supported ocamlfind version is now 1.8.1 (#15947 and #16046, fixes #14260 and #16015, by Pierre-Marie Pédrot and Théo Zimmermann).
Extraction¶
Changed:
ExtrOCamlInt63
no longer extractscomparison
toint
in OCaml; the extraction ofUint63.compare
andSint63.compare
was also adapted accordingly (#15294, fixes #15280, by Li-yao Xia).Changed: Extraction from
nat
to OCamlint
uses Stdlib instead of Pervasives (#15333, by Rudy Nicolo Peterson).Changed: The empty inductive type is now extracted to OCaml empty type available since OCaml 4.07 (#15967, by Pierre Roux).
Added: More extraction definitions for division and comparison of Z and N (#15098, by Li-yao Xia).
Fixed: Type
int
in filesNumber.v
,Decimal.v
andHexadecimal.v
have been renamed tosigned_int
(together with a compatibility aliasint
) so that they can be used in extraction without conflicting with OCaml'sint
type (#13460, fixes #7017 and #13288, by Hugo Herbelin).
Version 8.15¶
Summary of changes¶
Coq version 8.15 integrates many bug fixes, deprecations and cleanups as well as a few new features. We highlight some of the most impactful changes here:
The
apply with
tactic no longer renames arguments unless compatibility flagApply With Renaming
is set.Improvements to the
auto
tactic family, fixing theHint Unfold
behavior, and generalizing the use of discrimination nets.The
typeclasses eauto
tactic has a new best_effort option allowing it to return partial solutions to a proof search problem, depending on the mode declarations associated to each constraint. This mode is used by typeclass resolution during type inference to provide more precise error messages.Many commands and options were deprecated or removed after deprecation and more consistently support locality attributes.
The
Import
command is extended withimport_categories
to select the components of a module to import or not, including features such as hints, coercions, and notations.A visual Ltac debugger is now available in CoqIDE.
See the Changes in 8.15.0 section below for the detailed list of changes, including potentially breaking changes marked with Changed. Coq's reference manual for 8.15, documentation of the 8.15 standard library and developer documentation of the 8.15 ML API are also available.
Emilio Jesús Gallego Arias, Gaëtan Gilbert, Michael Soegtrop and Théo Zimmermann worked on maintaining and improving the continuous integration system and package building infrastructure.
Erik Martin-Dorel has maintained the Coq Docker images that are used in many Coq projects for continuous integration.
The OPAM repository for Coq packages has been maintained by Guillaume Claret, Karl Palmskog, Matthieu Sozeau and Enrico Tassi with contributions from many users. A list of packages is available at https://coq.inria.fr/opam/www/.
The Coq Platform has been maintained by Michael Soegtrop and Enrico Tassi.
Our current maintainers are Yves Bertot, Frédéric Besson, Ali Caglayan, Tej Chajed, Cyril Cohen, Pierre Corbineau, Pierre Courtieu, Maxime Dénès, Jim Fehrle, Julien Forest, Emilio Jesús Gallego Arias, Gaëtan Gilbert, Georges Gonthier, Benjamin Grégoire, Jason Gross, Hugo Herbelin, Vincent Laporte, Olivier Laurent, Assia Mahboubi, Kenji Maillard, Guillaume Melquiond, Pierre-Marie Pédrot, Clément Pit-Claudel, Pierre Roux, Kazuhiko Sakaguchi, Vincent Semeria, Michael Soegtrop, Arnaud Spiwack, Matthieu Sozeau, Enrico Tassi, Laurent Théry, Anton Trunov, Li-yao Xia and Théo Zimmermann. See the Coq Team face book page for more details.
The 41 contributors to this version are Tanaka Akira, Frédéric Besson, Juan Conejero, Ali Caglayan, Cyril Cohen, Adrian Dapprich, Maxime Dénès, Stéphane Desarzens, Christian Doczkal, Andrej Dudenhefner, Jim Fehrle, Emilio Jesús Gallego Arias, Attila Gáspár, Gaëtan Gilbert, Jason Gross, Hugo Herbelin, Jasper Hugunin, Bart Jacobs, Ralf Jung, Grant Jurgensen, Jan-Oliver Kaiser, Wojciech Karpiel, Fabian Kunze, Olivier Laurent, Yishuai Li, Erik Martin-Dorel, Guillaume Melquiond, Jean-Francois Monin, Pierre-Marie Pédrot, Rudy Peterson, Clément Pit-Claudel, Seth Poulsen, Pierre Roux, Takafumi Saikawa, Kazuhiko Sakaguchi, Michael Soegtrop, Matthieu Sozeau, Enrico Tassi, Laurent Théry, Anton Trunov and Théo Zimmerman.
The Coq community at large helped improve the design of this new version via the GitHub issue and pull request system, the Coq development mailing list coqdev@inria.fr, the coq-club@inria.fr mailing list, the Discourse forum and the Coq Zulip chat.
Version 8.15's development spanned 3 months from the release of Coq 8.14.0. Gaëtan Gilbert is the release manager of Coq 8.15. This release is the result of 384 merged PRs, closing 67 issues.
Changes in 8.15.0¶
Kernel¶
Fixed: Name clash in a computation of the type of parameters of functorial module types; this computation was provided for the purpose of clients using the algebraic form of module types such as
Print Module Type
(#15385, fixes #9555, by Hugo Herbelin).
Specification language, type inference¶
Changed:
Instance
warns about the default locality immediately rather than waiting until the instance is ready to be defined. This changes which command warns when the instance has a separate proof: theInstance
command itself warns instead of the proof closing command (such asDefined
). (#14705, by Gaëtan Gilbert).Removed: Arguments of section variables may no longer be renamed with
Arguments
(this was previously applied inconsistently) (#14573, by Gaëtan Gilbert).Added: Non-dependent implicit arguments can be provided explicitly using the syntax
(natural := term)
wherenatural
is the index of the implicit argument among all non-dependent arguments of the function, starting from 1 (#11099, by Hugo Herbelin).Added:
Succeed
, acontrol_command
that verifies that the givensentence
succeeds without changing the proof state (#14750, by Gaëtan Gilbert).Fixed: The
term.(qualid arg*)
syntax now takes into account the position of the main argumentterm
when computing the implicit arguments ofqualid
(#14606, fixes #4167, by Hugo Herbelin).Fixed: Source and target of coercions preserved by module instantiation (#14668, fixes #3527, by Hugo Herbelin).
Fixed: Made reference manual consistent with the implementation regarding the role of recursively non-uniform parameters of inductive types in the nested positivity condition (#14967, fixes #14938, by Hugo Herbelin)
Notations¶
Changed: Terms printed in error messages may be more verbose if syntactic sugar would make it appear that the obtained and expected terms only differ in existential variables (#14672, by Gaëtan Gilbert).
Removed: the
Numeral Notation
command that was renamed toNumber Notation
in 8.13. (#14819, by Pierre Roux).Removed: primitive float notations
<
,<=
and==
that were replaced by<?
,<=?
and=?
in 8.13. (#14819, by Pierre Roux).Removed: primitive integer notations
\%
,<
,<=
and==
that were replaced bymod
,<?
,<=?
and=?
in 8.13. (#14819, by Pierre Roux).Added: Include floats in the number notation mechanism (#14525, by Pierre Roux).
Added: Coercion entries and
ident
/global
entries in custom notations now respect theonly parsing
modifier (#15340, fixes #15335, by Hugo Herbelin).Fixed:
Reserved Infix
now accept further parameters in the infix notation (#14379, fixes #11402, by Hugo Herbelin).Fixed: Useless self reference when printing abbreviations declared in nested modules (#14493, fixes one part of #12777 and #14486, by Hugo Herbelin).
Fixed: anomalies with notation applied in
match
patterns when the notation have a notation variable at head (#14713, fixes #14708, by Hugo Herbelin).Fixed: Regression in parsing error reporting in case of empty custom entry (#15338, fixes #15334, by Hugo Herbelin).
Tactics¶
Changed:
apply with
does not rename arguments unless using compatibility flagApply With Renaming
(#13837, fixes #13759, by Gaëtan Gilbert).Porting hint: if the renaming is because of a goal variable (eg
intros x; apply foo with (x0 := bar)
whereAbout foo.
says the argument is calledx
) it is probably caused by an interaction with implicit arguments andapply @foo with (x := bar)
will usually be a backwards compatible fix.Changed:
Hint Unfold
in discriminated databases now respects its specification, namely that a constant may be unfolded only when it is the head of the goal. The previous behavior was to perform unfolding on any goal, without any limitation.An unexpected side-effect of this was that a database that contained
Unfold
hints would sometimes trigger silent strong βι-normalization of the goal. Indeed,unfold
performs such a normalization regardless of the presence of its argument in the goal. This does introduce a bit of backwards incompatibility, but it occurs in very specific situations and is easily circumvented. Since by default hint bases are not discriminated, it means that incompatibilities are typically observed when adding unfold hints to the typeclass database.In order to recover the previous behavior, it is enough to replace instances of
Hint Unfold foo.
withHint Extern 4 => progress (unfold foo).
. A less compatible but finer-grained change can be achieved by only adding the missing normalization phase withHint Extern 4 => progress (lazy beta iota).
(#14679, fixes #14874, by Pierre-Marie Pédrot).Changed: Correctly consider variables without a body to be rigid for the pattern recognition algorithm of discriminated hints (#14722, by Pierre-Marie Pédrot).
Changed: Use discrimination nets for goals containing evars in all
auto
tactics. It essentially makes the behavior of undiscriminated databases to be the one of discriminated databases where all constants are considered transparent. This may be incompatible with previous behavior in very rare cases (#14848, by Pierre-Marie Pédrot).Changed: The
choice
strategy forrewrite_strat
is now of arbitrary arity (#14989, fixes #6109, by Gaëtan Gilbert).Changed: The
exact
tactic now takes auconstr
as argument instead of an ad-hoc one. In very rare cases, this can change the order of resolution of dependent evars when used over several goals at once (#15171, by Pierre-Marie Pédrot).Changed:
cbn
interprets the combination of the!
and/
modifiers (fromArguments
) to mean "unfold as soon as all arguments before the/
are provided and all arguments marked with!
reduce to a constructor". This makes it unfold more often than without the/
when all arguments are provided. Previously adding/
would only prevent unfolding when insufficient arguments are provided without adding new unfoldings.Note that this change only takes effect in default mode (as opposed to when
simpl nomatch
was used) (#15204, fixes #4555 and #7674, by Gaëtan Gilbert).Removed: the deprecated new auto tactic (#14527, by Pierre-Marie Pédrot).
Removed: deprecated syntax for
instantiate
using capitalizedValue
orType
(#15193, by Gaëtan Gilbert).Removed: deprecated
autoapply ... using
syntax forautoapply
(#15194, by Gaëtan Gilbert).Deprecated: the
bfs eauto
tactic. Since its introduction it has behaved exactly like theeauto
tactic. Usetypeclasses eauto
with thebfs
flag instead (#15314, fixes #15300, by Pierre-Marie Pédrot).Added: The
zify
tactic can now recognizePos.Nsucc_double
,Pos.Ndouble
,N.succ_double
,N.double
,N.succ_pos
,N.div2
,N.pow
,N.square
, andZ.to_pos
. Moreover, importing moduleZifyBool
lets it recognizePos.eqb
,Pos.leb
,Pos.ltb
,N.eqb
,N.leb
, andN.ltb
(#10998, by Kazuhiko Sakaguchi).Added: best_effort option to
typeclasses eauto
, to return a partial solution to its initial proof-search problem. The goals that can remain unsolved are determined according to the modes declared for their head (seeHint Mode
). This is used by typeclass resolution during type inference to provide more informative error messages. (#13952, fixes #13942 and #14125, by Matthieu Sozeau).Added: A new
Keep Equalities
table to selectively control the preservation of subterm equalities for theinjection
tactic. It allows a finer control than the boolean flagKeep Proof Equalities
that acts globally. (#14439, by Pierre-Marie Pédrot).Added:
simple congruence
tactic which works likecongruence
but does not unfold definitions. (#14657, fixes #13778 and #5394 and #13189, by Andrej Dudenhefner).Added: Small enhancement of unification in the presence of local definitions (#14673, fixes #4415, by Hugo Herbelin).
Added:
dfs
option intypeclasses eauto
to use depth-first search (#14693, fixes #13859, by Ali Caglayan).Fixed: More flexible hypothesis specialization in
congruence
. (#14650, fixes #14651 and #14662, by Andrej Dudenhefner).Fixed: Added caching to congruence initialization to avoid quadratic runtime. (#14683, fixes #5548, by Andrej Dudenhefner).
Fixed: Correctly handle matching up to η-expansion in discriminated hints (#14732, fixes #14731, by Pierre-Marie Pédrot).
Fixed: Old unification understands some inductive cumulativity (#14758, fixes #14734 and #6976, by Gaëtan Gilbert).
Fixed: The
clear dependent
tactic now does not backtrack internally, preventing an exponential blowup (#14984, fixes #11689, by Pierre-Marie Pédrot).Fixed:
setoid_rewrite
now works when the rewriting lemma has non dependent arguments and rewriting under binders (#14986, fixes #5369, by Gaëtan Gilbert).Fixed: Regression in 8.14.0 and 8.14.1 with action pattern
%
inas
clause of tacticspecialize
(#15245, fixes #15244, by Hugo Herbelin).
SSReflect¶
Changed: rewrite generates subgoals in the expected order (side conditions first, by default) also when rewriting with a setoid relation. (#14314, fixes #5706, by Enrico Tassi).
Removed: The ssrsearch plugin and the ssr Search command (#13760, by Jim Fehrle).
Added: port the additions made to
ssrbool.v
in math-comp PR #757, namelyreflect
combinatorsnegPP
,orPP
,andPP
andimplyPP
(#15059, by Christian Doczkal).Fixed: SSR patterns now work with primitive values such as ints, floats or arrays (#14660, fixes #12770, by Juan Conejero).
Fixed: A bug where
suff
would fail due to use ofapply
under the hood. (#14687, fixes #14678, by Ali Caglayan helped by Enrico Tassi).
Commands and options¶
Changed:
About
andPrint
now display all known argument names (#14596, grants #13830, by Hugo Herbelin).Changed:
Typeclasses Transparent
andTypeclasses Opaque
support#[local]
,#[export]
and#[global]
attributes (#14685, fixes #14513, by Gaëtan Gilbert).Changed: In extraction to OCaml, empty types in
Type
(such asEmpty_set
) are now extracted to an abstract type (empty by construction) rather than to the OCaml'sunit
type (#14802, fixes a remark at #14801, by Hugo Herbelin).Changed: Closed modules now live in a separate namespace from open modules and sections. (#15078, fixes #14529, by Gaëtan Gilbert).
Removed: boolean attributes
monomorphic
,noncumulative
andnotemplate
that were replaced bypolymorphic=no
,cumulative=no
andtemplate=no
in 8.13. (#14819, by Pierre Roux).Removed: command
Grab Existential Variables
that was deprecated in 8.13. UseUnshelve
that is mostly equivalent, up to the reverse order of the resulting subgoals. (#14819, by Pierre Roux).Removed: command
Existential
that was deprecated in 8.13. UseUnshelve
andexact
. (#14819, by Pierre Roux).Removed: the
-outputstate
command line argument and the corresponding vernacular commandsWrite State
andRestore State
(#14940, by Pierre-Marie Pédrot)Deprecated: ambiguous
Proof using
andCollection
usage (#15056, fixes #13296, by Wojciech Karpiel).Deprecated:
Universal Lemma Under Conjunction
flag that was introduced for compatibility with Coq versions prior to 8.4 (#15272, by Théo Zimmermann).Deprecated: using
Hint Cut
,Hint Mode
,Hint Transparent
,Hint Opaque
,Typeclasses Transparent
orTypeclasses Opaque
without an explicit locality outside sections. (#14697, by Pierre-Marie Pédrot, and #14685, by Gaëtan Gilbert)Added: The
Mangle Names Light
flag, which changes the behavior ofMangle Names
. For example, the namefoo
becomes_0
withMangle Names
, but withMangle Names Light
set, it will become_foo
. (#14695, fixes #14548, by Ali Caglayan).Added: The
Hint Cut
,Hint Mode
,Hint Transparent
,Hint Opaque
,Typeclasses Transparent
andTypeclasses Opaque
commands now accept thelocal
,export
andglobal
locality attributes inside sections. With either attribute, the commands will trigger thenon-local-section-hint
warning if the arguments refer to local section variables (#14697, by Pierre-Marie Pédrot, and #14685, fixes #14513, by Gaëtan Gilbert).Added:
projections(primitive)
attribute to make a record use primitive projections (#14699, fixes #13150, by Ali Caglayan).Added: Syntax for
import_categories
providing selective import of module components (egImport(notations) M
(#14892, by Gaëtan Gilbert).Added:
Search
understands modifierin
as an alias ofinside
(#15139, fixes #14930, by Gaëtan Gilbert). This is intended to ease transition for ssreflect Search users.Fixed: interaction of Program's obligation state and modules and sections: obligations started in a parent module or section are not available to be solved until the submodules and subsections are closed (#14780, fixes #14446, by Gaëtan Gilbert).
Fixed:
Eval
andCompute
now beta-iota-simplify the type of the result, likeCheck
does (#14901, fixes #14899, by Hugo Herbelin)
Command-line tools¶
Changed: Coqdoc options
--coqlib
and--coqlib_path
have been renamed to--coqlib_url
and--coqlib
to make them more consistent with flags used by other Coq executables (#14059, by Emilio Jesus Gallego Arias).Changed: Syntax of
_CoqProject
files:-arg
is now handled by coq_makefile and not bymake
. Unquoted#
now start line comments. (#14558, by Stéphane Desarzens, with help from Jim Fehrle and Enrico Tassi).Changed:
Require
now selects files whose logical name exactly matches the required name, making it possible to unambiguously select a given file: if several-Q
or-R
options bind the same logical name to a different file, the option appearing last on the command line takes precedence. Moreover, it is now an error to require a file using a partial logical name which does not resolve to a non-ambiguous path (#14718, by Hugo Herbelin).Changed:
coq_makefile
now declares variableCOQBIN
to avoid warnings inmake --warn
mode (#14787, by Clément Pit-Claudel).Changed:
coqchk
respects theKernel Term Sharing
flag instead of forcing it on (#14957, by Gaëtan Gilbert)Removed: These options of coq_makefile:
-extra
,-extra-phony
,-custom
,-no-install
,-install
,-no-opt
,-byte
. Support for subdirectories is also removed. (#14558, by Stéphane Desarzens, with help from Jim Fehrle and Enrico Tassi).Added: coq_makefile now takes the
-docroot
option as alternative to theINSTALLCOQDOCROOT
variable (#14558, by Stéphane Desarzens, with help from Jim Fehrle and Enrico Tassi).Fixed: Various
coqdep
issues with theFrom
clause ofRequire
and a few inconsistencies betweencoqdep
andcoqc
disambiguation ofRequire
(#14718, fixes #11631 and #14539, by Hugo Herbelin).Fixed:
coq_makefile
has improved logic when dealing with incorrect_CoqProject
files (#13541, fixes #9319, by Fabian Kunze).Fixed:
coqdep
was confusing periods occurring in comments with periods ending Coq sentences (#14996, fixes #7393, by Hugo Herbelin).
CoqIDE¶
Changed: CoqIDE unicode keys for brackets (e.g.
langle
) now bind to unicode mathematical symbols rather than unicode CJK brackets (#14452, by Bart Jacobs).Changed: All occurrences of the name
CoqIde
toCoqIDE
. This may cause issues with installing and uninstalling desktop icons, causing apparent duplicates. (#14696, fixes #14310, by Ali Caglayan).Added: Initial version of a visual debugger in CoqIDE. Supports setting breakpoints visually and jumping to the stopping point plus continue, step over, step in and step out operations. Displays the call stack and variable values for each stack frame. Currently only for Ltac. See the documentation here. (#14644, fixes #13967, by Jim Fehrle)
Fixed: It is now possible to deactivate the unicode completion mechanism in CoqIDE (#14863, by Pierre-Marie Pédrot).
Standard library¶
Changed: Permutation-related Proper instances are now at default priority instead of priority
10
(#14574, fixes #14571, by Gaëtan Gilbert).Changed: The new type of
epsilon_smallest
is(exists n : nat, P n) -> { n : nat | P n /forall k, P k -> n <= k }
. Here the minimality ofn
is expressed byforall k, P k -> n <= k
corresponding to the intuitive meaning of minimality "the others are greater", whereas the previous version used the negative equivalent formulationforall k, k < n -> ~P k
. Scripts usingepsilon_smallest
can easily be adapted using lemmasle_not_lt
andlt_not_le
from the standard library. (#14601, by Jean-Francois Monin).Changed:
ltb
andleb
functions forascii
, into comparison-based definition (#14234, by Yishuai Li).Removed: the file
Numeral.v
that was replaced byNumber.v
in 8.13. (#14819, by Pierre Roux).Removed: some
*_invol
functions that were renamed*_involutive
for consistency with the remaining of the stdlib in 8.13. (#14819, by Pierre Roux).Deprecated:
frexp
andldexp
inFloatOps.v
, renamedZ.frexp
andZ.ldexp
(#15085, by Pierre Roux).Added: A proof that incoherent equivalences can be adjusted to adjoint equivalences in
Logic.Adjointification
. (#13408, by Jasper Hugunin).Added:
ltb
andleb
functions forstring
, and some lemmas about them;Added: simple non dependent product
slexprod
inRelations/Relation_Operators.v
and its proof of well-foundnesswf_slexprod
inWellfounded/Lexicographic_Product.v
(#14809, by Laurent Thery).Added: The notations
(x; y)
,x.1
,x.2
for sigT are now exported and available afterImport SigTNotations.
(#14813, by Laurent Théry).Added: The function
sigT_of_prod
turns a pairA * B
into{_ : A & B}
. Its inverse function isprod_of_sigT
. This is shown by theoremssigT_prod_sigT
andprod_sigT_prod
. (#14813, by Laurent Théry).Fixed:
split_combine
lemma for lists, making it usable (#14458, by Yishuai Li).
Infrastructure and dependencies¶
Changed: Coq's continuous integration now provides a more accessible Windows installer artifact in the "Checks" GitHub tab, both for pull requests and the
master
branch.This facilitates testing Coq's bleeding edge builds on Windows, and should be more reliable than the previous setup (#12425, by Emilio Jesus Gallego Arias).
Changed: Coq's
./configure
script has gone through a major cleanup. In particular, the following options have been removed:-force-caml-version
,-force-findlib-version
: Coq won't compile with OCaml or findlib lower than the required versions;-vmbyteflags
,-custom
,-no-custom
: linking options for toplevels are now controlled intopbin/dune
;-ocamlfind
: Coq will now use the toolchain specified in the Dune configuration; this can be controlled using the workspaces feature;-nodebug
: Coq will now follow the standard, which is to always pass-g
to OCaml; this can be modified using a custom Dune workspace;-flambda-opts
: compilation options are now set in Coq's rootdune
file, can be updated using a custom Dune workspace;-local
,-bindir
,-coqdocdir
,-annotate
,-camldir
,-profiling
: these flags were deprecated in 8.14, and are now removed.
Moreover, the
-annot
and-bin-annot
flags only take effect to setcoq-makefile
's defaults. (#14189, by Emilio Jesus Gallego Arias).Changed: Configure will now detect the Dune version, and will correctly pass
-etcdir
and-docdir
to the install procedure if Dune >= 2.9 is available. Note that the-docdir
configure option now refers to root path for documentation. If you would like to install Coq documentation infoo/coq
, use-docdir foo
. (#14844, by Emilio Jesus Gallego Arias).Changed: OCaml 4.13 is now officially supported (#14879, by Emilio Jesus Gallego Arias)
Changed: Sphinx 3.0.2 or above is now required to build the reference manual (#14963, by Théo Zimmermann)
Extraction¶
Changed: replaced
Big
module withBig_int_Z
functions fromzarith
.OCaml code extracted with the following modules should be linked to the Zarith library.
ExtrOcamlNatBigInt
ExtrOcamlZBigInt
Removed
ExtrOcamlBigIntConv
module.(#8252, by Yishuai Li).
Fixed: compilation errors in ExtrOcamlString and ExtrOcamlNativeString (#15075, fixes #15076, by Yishuai Li).
Changes in 8.15.1¶
Kernel¶
Fixed: cases of incompletenesses in the guard condition for fixpoints in the presence of cofixpoints or primitive projections (#15498, fixes #15451, by Hugo Herbelin).
Fixed: inconsistency when using module subtyping with squashed inductives (#15839, fixes #15838 (reported by Pierre-Marie Pédrot), by Gaëtan Gilbert).
Tactics¶
Fixed:
rewrite_strat
regression in 8.15.0 related toTransitive
instances (#15577, fixes #15568, by Gaëtan Gilbert).Fixed: When
setoid_rewrite
succeeds in rewriting at some occurrence but the resulting equality is the identity, it now tries rewriting in subterms of that occurrence instead of giving up (#15612, fixes #8080, by Gaëtan Gilbert).Fixed: Ill-typed goals created by
clearbody
in the presence of transitive dependencies in the body of a hypothesis (#15634, fixes #15606, by Hugo Herbelin).Fixed:
cbn
knows to refold fixpoints whenArguments
with/
and!
was used (#15653, fixes #15567, by Gaëtan Gilbert).
Command-line tools¶
Fixed: a bug where
coqc -vok
was not creating an empty '.vok' file. (#15745, by Ramkumar Ramachandra).
CoqIDE¶
Fixed: Line numbers shown in the Errors panel were incorrect; they didn't match the error locations in the script panel (#15532, fixes #15531, by Jim Fehrle).
Fixed: anomaly when using proof diffs with no focused goal (#15633, fixes #15578, by Jim Fehrle).
Fixed: Attempted edits to the processed part of a buffer while Coq is busy processing a request are now ignored to ensure "processed" highlighting is accurate (#15714, fixes #15733 and #15675 and #15725, by Jim Fehrle).
Changes in 8.15.2¶
Tactics¶
Added:
intuition
anddintuition
useTauto.intuition_solver
(defined asauto with *
) instead of hardcodingauto with *
. This makes it possible to change the default solver withLtac Tauto.intuition_solver ::= ...
(#15866, fixes #7725, by Gaëtan Gilbert).Fixed: uncaught exception
UnableToUnify
with bidirectionality hints (#16066, fixes #16063, by Gaëtan Gilbert).
Version 8.14¶
Summary of changes¶
Coq version 8.14 integrates many usability improvements, as well as an important change in the core language. The main changes include:
The internal representation of
match
has changed to a more space-efficient and cleaner structure, allowing the fix of a completeness issue with cumulative inductive types in the type-checker. The internal representation is now closer to the user-level view ofmatch
, where the argument context of branches and the inductive bindersin
andas
do not carry type annotations.A new
coqnative
binary performs separate native compilation of libraries, starting from a.vo
file. It is supported bycoq_makefile
.Improvements to typeclasses and canonical structure resolution, allowing more terms to be considered as classes or keys.
More control over notations declarations and support for primitive types in string and number notations.
Removal of deprecated tactics, notably
omega
, which has been replaced by a greatly improvedlia
, along with many bug fixes.New Ltac2 APIs for interaction with Ltac1, manipulation of inductive types and printing.
Many changes and additions to the standard library in the numbers, vectors and lists libraries. A new signed primitive integers library
Sint63
is available in addition to the unsignedUint63
library.
See the Changes in 8.14.0 section below for the detailed list of changes, including potentially breaking changes marked with Changed. Coq's reference manual, documentation of the standard library and developer documentation of the ML API are also available.
Emilio Jesús Gallego Arias, Gaëtan Gilbert, Michael Soegtrop and Théo Zimmermann worked on maintaining and improving the continuous integration system and package building infrastructure.
Erik Martin-Dorel has maintained the Coq Docker images that are used in many Coq projects for continuous integration.
The OPAM repository for Coq packages has been maintained by Guillaume Claret, Karl Palmskog, Matthieu Sozeau and Enrico Tassi with contributions from many users. A list of packages is available at https://coq.inria.fr/opam/www/.
The Coq Platform has been maintained by Michael Soegtrop and Enrico Tassi.
Our current maintainers are Yves Bertot, Frédéric Besson, Ali Caglayan, Tej Chajed, Cyril Cohen, Pierre Corbineau, Pierre Courtieu, Maxime Dénès, Jim Fehrle, Julien Forest, Emilio Jesús Gallego Arias, Gaëtan Gilbert, Georges Gonthier, Benjamin Grégoire, Jason Gross, Hugo Herbelin, Vincent Laporte, Olivier Laurent, Assia Mahboubi, Kenji Maillard, Guillaume Melquiond, Pierre-Marie Pédrot, Clément Pit-Claudel, Pierre Roux, Kazuhiko Sakaguchi, Vincent Semeria, Michael Soegtrop, Arnaud Spiwack, Matthieu Sozeau, Enrico Tassi, Laurent Théry, Anton Trunov, Li-yao Xia and Théo Zimmermann. See the Coq Team face book page for more details.
The 54 contributors to this version are Reynald Affeldt, Arthur Azevedo de Amorim, Yves Bertot, Frédéric Besson, Lasse Blaauwbroek, Ana Borges, Ali Caglayan, Cyril Cohen, Pierre Courtieu, Maxime Dénès, Stéphane Desarzens, Andrej Dudenhefner, Jim Fehrle, Yannick Forster, Simon Friis Vindum, Gaëtan Gilbert, Jason Gross, Samuel Gruetter, Stefan Haan, Hugo Herbelin, Jasper Hugunin, Emilio Jesús Gallego Arias, Jacques-Henri Jourdan, Ralf Jung, Jan-Oliver Kaiser, Fabian Kunze, Vincent Laporte, Olivier Laurent, Yishuai Li, Barry M. Trager, Kenji Maillard, Erik Martin-Dorel, Guillaume Melquiond, Isaac Oscar Gariano, Pierre-Marie Pédrot, Rudy Peterson, Clément Pit-Claudel, Pierre Roux, Takafumi Saikawa, Kazuhiko Sakaguchi, Gabriel Scherer, Vincent Semeria, shenlebantongying, Avi Shinnar, slrnsc, Michael Soegtrop, Matthieu Sozeau, Enrico Tassi, Hendrik Tews, Anton Trunov, Karolin Varner, Li-yao Xia, Beta Ziliani and Théo Zimmermann.
The Coq community at large helped improve the design of this new version via the GitHub issue and pull request system, the Coq development mailing list coqdev@inria.fr, the coq-club@inria.fr mailing list, the Discourse forum and the Coq Zulip chat.
Version 8.14's development spanned 9 months from the release of Coq 8.13.0. Guillaume Melquiond is the release manager of Coq 8.14. This release is the result of 522 merged PRs, closing ~150 issues.
Changes in 8.14.0¶
Kernel¶
Changed: The term representation of pattern-matchings now uses a compact form that provides a few static guarantees such as eta-expansion of branches and return clauses and is usually more efficient. The most visible user change is that for the time being, the
destruct
tactic and its variants generate dummy cuts (β redexes) in the branches of the generated proof. This can also generate very uncommon backwards incompatibilities, such as a change of occurrence numbering for subterms, or breakage of unification in complex situations involving pattern-matchings whose underlying inductive type declares let-bindings in parameters, arity or constructor types. For ML plugin developers, an in-depth description of the new representation, as well as porting tips, can be found in dev/doc/case-repr.md (#13563, fixes #3166, by Pierre-Marie Pédrot).Changed: Linking of native-code libraries used by
native_compute
is now delayed until an actual call to thenative_compute
machinery is performed. This should make Coq more responsive on some systems (#13853, fixes #13849, by Guillaume Melquiond).Removed: The ability to change typing flags inside sections to prevent exploiting a weakness in
Print Assumptions
(#14395, fixes #14317, by Gaëtan Gilbert).
Specification language, type inference¶
Changed: The hints mode
!
matches a term iff the applicative head is not an existential variable. It now also matches projections applied to any term or amatch
on any term. (#14392, by Matthieu Sozeau).Removed: The little used
:>
type cast, which was only interpreted in Program-mode (#13911, by Jim Fehrle and Théo Zimmermann).Added: Enable canonical
fun _ => _
projections, see Canonical Structures for details. (#14041, by Jan-Oliver Kaiser and Pierre Roux, reviewed by Cyril Cohen and Enrico Tassi).Added:
Canonical Structure
declarations now accept dependent function typesforall _, _
as keys. (#14386, by Jan-Oliver Kaiser and Kazuhiko Sakaguchi).Added: Ability to declare primitive projections as class, for dependent typeclass resolutions (#9711, fixes #12975, by Matthieu Sozeau).
Fixed: Multiple printing of same warning about unused variables catching several cases (#14261, fixes #14207, by Hugo Herbelin).
Fixed: Constants
id
andnot
were unduly set opaque in some parts of the unification algorithm (#14371, fixes #14374, by Hugo Herbelin).
Notations¶
Changed: Flag
Printing Notations
no longer controls whether strings and numbers are printed raw (#13840, by Enrico Tassi).Changed: The error
Argument X was previously inferred to be in scope XXX_scope but is here used in YYY_scope.
is now the warning[inconsistent-scopes,syntax]
and can be silenced by specifying the scope of the argument (#13965, by Enrico Tassi).Removed: Decimal-only number notations which were deprecated in 8.12. (#13842, by Pierre Roux).
Added:
Number Notation
andString Notation
now support parsing and printing of primitive floats, primitive arrays and type constants of primitive types. (#13519, fixes #13484 and #13517, by Fabian Kunze, with help of Jason Gross)Added: Flag
Printing Raw Literals
to control whether strings and numbers are printed raw. (#13840, by Enrico Tassi).Added: Let the user specify a scope for abbreviation arguments, e.g.
Notation abbr X := t (X in scope my_scope)
. (#13965, by Enrico Tassi).Added: Look-ahead of tokens is changed from sequential to tree-based, allowing more automatic rule factorizations in notations (#14070, by Hugo Herbelin).
Fixed: Non-local custom entries survive module closing and are declared when a file is Required (#14183, fixes #13654, by Gaëtan Gilbert).
Fixed:
ident
modifier in custom entry notations gave fatal errors at printing time (#14257, fixes #14211, by Hugo Herbelin).Fixed: Anomaly when overriding a notation with different applicability in
match
patterns (#14377, fixes #13966, by Hugo Herbelin).
Tactics¶
Changed: More systematic checks that occurrences of an
at
clause are valid in tactics such asrewrite
orpattern
(#13568, fixes #13566, by Hugo Herbelin).Removed:
fail
andgfail
, which formerly accepted negative values as a parameter, now give syntax errors for negative values (#13469, by Jim Fehrle).Removed: Deprecated flag
Bracketing Last Introduction Pattern
affecting the behavior of trailing disjunctive introduction patterns is definitively removed (#13509, by Hugo Herbelin).Removed: The
omega
tactic (deprecated in 8.12) and four* Omega *
flags. Uselia
instead. (#13741, by Jim Fehrle, who addressed the final details, building on much work by Frédéric Besson, who greatly improvedlia
, Maxime Dénès, Vincent Laporte and with the help of many package maintainers, among others).Removed: convert_concl_no_check. Use
change_no_check
instead (#13761, by Jim Fehrle).Removed: double induction tactic. Replace
double induction ident ident
withinduction ident; induction ident
(orinduction ident ; destruct ident
depending on the exact needs). Replacedouble induction natural1 natural2
withinduction natural1; induction natural3
wherenatural3
is the result ofnatural2 - natural1
(#13762, by Jim Fehrle).Deprecated: In
change
andchange_no_check
, theat ... with ...
form is deprecated. Usewith ... at ...
instead. Forat ... with ... in H |-
, usewith ... in H at ... |-
. (#13696, by Jim Fehrle).Deprecated: The micromega option
Simplex
, which is currently set by default (#13781, by Frédéric Besson).Deprecated: the undocumented
new auto
tactic (#14528, by Pierre-Marie Pédrot).Added:
lia
supports the boolean operatorBool.implb
(#13715, by Frédéric Besson).Added:
zify
(lia
/nia
) support fordiv
,mod
,pow
forNat
(viaZifyNat
module) andN
(viaZifyN
module). The signature ofZ_div_mod_eq_full
has no assumptions. (#14037, fixes #11447, by Andrej Dudenhefner, Jason Gross, and Frédéric Besson).Added: Ltac2 now has a
unify
tactic (#14089, fixes #14083, by Samuel Gruetter).Added:
inversion_sigma
can now be applied to a specified hypothesis and additionally supports intropatterns, so it can be used much likeinduction
andinversion
. Additionally,inversion_sigma
now supports the typesex
(exists x : A, P x
) andex2
(exists2 x : A, P x & Q x
) in cases where the first argumentA
is aProp
(#14174, by Jason Gross).Added:
zify
(lia
/nia
) support forSint63
. (#14408, by Ana Borges, with help from Frédéric Besson).Fixed: Possible collision between a user-level name and an internal name when using the
%
introduction pattern (#13512, fixes #13413, by Hugo Herbelin).Fixed:
simpl
andhnf
now reduce primitive functions on primitive integers, floats and arrays (#13699, fixes #13579, by Pierre Roux).Fixed: Setoid rewriting now remembers the (invisible) binder names of non-dependent product types. SSReflect's rewrite tactic expects these names to be retained when using
rewrite foo in H
. This also fixes SSRrewrite foo in H *
erroneously revertingH
. (#13882, fixes #12011, by Gaëtan Gilbert).Fixed: Properly expand projection parameters in hint discrimination nets. (#14033, fixes #9000, #14009, by Pierre-Marie Pédrot).
Fixed: anomalies caused by empty strings in Ltac notations are now errors (#14378, fixes #14124, by Hugo Herbelin).
Fixed: Print a message instead of a Diff_Failure anomaly when old and new goals can't be matched; show the goal without diff highlights (#14457, fixes #14425, by Jim Fehrle).
Fixed: Anomaly of
destruct
on terms with dependent variables unused in goal (#15099, fixes #11504 and #14090, by Lasse Blaauwbroek and Hugo Herbelin).Fixed: Correct convertibility of multiple terms selected by patterns in tactics such as
set
when these terms have subterms inSProp
(#14610, fixes #14609, by Hugo Herbelin).
Tactic language¶
Changed: Renamed Ltac2
Bool.eq
intoBool.equal
for uniformity. The old function is now a deprecated alias (#14128, by Pierre-Marie Pédrot).Added: A
printf
macro to Ltac2. It can be made accessible by importing theLtac2.Printf
module. See the documentation there for more information (#13236, fixes #10108, by Pierre-Marie Pédrot).Added: A function
Ltac1.lambda
allowing to embed Ltac2 functions into Ltac1 runtime values (#13442, fixes #12871, by Pierre-Marie Pédrot).Added: Ltac2 commands defining terms now accept the
deprecated
attribute (#13774, fixes #12317, by Pierre-Marie Pédrot).Added: Allow the presence of type casts for function return values, let bindings and global definitions in Ltac2 (#13914, by Pierre-Marie Pédrot).
Added: The Ltac2 API
Ltac2.Ind
for manipulating inductive types (#13920, fixes #10095, by Pierre-Marie Pédrot).Added: Allow scope delimiters in Ltac2
open_constr:(...)
quotation (#13939, fixes #12806, by Pierre-Marie Pédrot).Added: A FFI to convert between Ltac1 and Ltac2 identifiers (#13997, fixes #13996, by Pierre-Marie Pédrot).
Added: Lazy evaluating boolean operators
lazy_and
,lazy_or
,lazy_impl
and infix notations&&
and||
to the Ltac2Bool.v
library l. (#14081, fixes #13964, by Michael Soegtrop).Fixed: Ltac2 notations now correctly take into account their assigned level (#14094, fixes #11866, by Pierre-Marie Pédrot).
Commands and options¶
Changed:
Hint Rewrite
now supports locality attributes (includingexport
) like other Hint commands (#13725, fixes #13724, by Gaëtan Gilbert).Changed: In
Record
, alpha-rename the variable associated with the record to avoid alpha-renaming parameters of projections (#13852, fixes #13727, by Li-yao Xia).Changed: Improve the
Coercion
command to reduce the number of ambiguous paths to report. A pair of multiple inheritance paths that can be reduced to smaller adjoining pairs will not be reported as ambiguous paths anymore. (#13909, by Kazuhiko Sakaguchi).Changed: The printing order of
Print Classes
andPrint Graph
, due to the changes for the internal tables of coercion classes and coercion paths. (#13912, by Kazuhiko Sakaguchi).Removed: The Hide Obligations flag, deprecated in 8.12 (#13758, by Jim Fehrle).
Removed: SearchHead command. Use the
headconcl:
clause ofSearch
instead (#13763, by Jim Fehrle).Removed:
Show Zify Spec
,Add InjTyp
and 11 similarAdd *
commands. ForShow Zify Spec
, useShow Zify UnOpSpec
orShow Zify BinOpSpec
instead. ForAdd *
,Use Add Zify *
intead ofAdd *
(#13764, by Jim Fehrle).Deprecated: Like hints, typeclass instances added outside of sections without an explicit locality now generate a deprecation warning. See Hint (#14208, fixes #13562, by Pierre-Marie Pédrot).
Deprecated: the
Regular Subst Tactic
flag (#14336, by Pierre-Marie Pédrot).Added:
Debug
to control debug messages, functioning similarly to the warning system (#13202, by Maxime Dénès and Gaëtan Gilbert). The following flags have been converted (such thatSet Flag
becomesSet Debug "flag"
):Debug Unification
tounification
Debug HO Unification
toho-unification
Debug Tactic Unification
totactic-unification
Congruence Verbose
tocongruence
Debug Cbv
tocbv
Debug RAKAM
toRAKAM
Debug Ssreflect
tossreflect
Added: The Ltac2 grammar can now be printed using the Print Grammar ltac2 command (#14093, fixes #14092, by Pierre-Marie Pédrot).
Added:
Instance
now accepts theexport
locality attribute (#14148, by Pierre-Marie Pédrot).Fixed: extraction failure of a parameterized type in
Prop
exported in an module interface as an assumption inType
(#14102, fixes #14100, by Hugo Herbelin).Fixed: Print Assumptions now treats delayed opaque proofs generated by vos compilation as if they were axioms (#14382, fixes #13589, by Pierre-Marie Pédrot).
Fixed: Incorrect de Bruijn index handling in vernac class declaration, preventing users from marking existing instances of existing classes which are primitive projections (#14664, fixes #14652, by Ali Caglayan and Hugo Herbelin).
Command-line tools¶
Changed:
coqc
now enforces that at most a single.v
file can be passed in the command line. Support for multiple.v
files in the form ofcoqc f1.v f2.v
didn't properly work in 8.13, tho it was accepted. (#13876, by Emilio Jesus Gallego Arias).Changed:
coqdep
now reports an error if files specified on the command line don't exist or if it encounters unreadable files. Unknown options now generate a warning. Previously these conditions were ignored. (#14024, fixes #14023, by Hendrik Tews).Changed: Makefiles produced by
coq_makefile
now use.DELETE_ON_ERROR
(#14238, by Gaëtan Gilbert).Removed: Previously deprecated command line options
-sprop-cumulative
and-input-state
and its alias-is
(#13822, by Gaëtan Gilbert).Added:
coq_makefile
-madeMakefile
s now support inclusion of a.local-late
file at the end, allowing the user to access more variables (#12411, fixes #10912, by Jason Gross).Fixed: Failure of extraction in the presence of inductive types with local definitions in parameters (#13624, fixes #13581, by Hugo Herbelin).
Fixed: File name was missing in coqdoc error position reporting (#14285, fixes #14283, by Arthur Charguéraud and Hugo Herbelin).
Native Compilation¶
Changed:
coq_makefile
now uses thecoqnative
binary to generate native compilation files. Project files also understand directly the-native-compiler
flag without having to wrap it with-arg
(#14265, by Pierre-Marie Pédrot).Deprecated: the
-native-compiler
option for coqc. It is now recommended to use the Split compilation of native computation files binary instead to generate native compilation files ahead of time (#14309, by Pierre-Marie Pédrot).Added: A standalone
coqnative
binary that performs native compilation out ofvo
files, allowing to split library compilation from native compilation. See Split compilation of native computation files. The hybrid build system was adapted to perform a split compilation on the stdlib (#13287, by Pierre-Marie Pédrot).
CoqIDE¶
Added: Ltac debugger support in CoqIDE (see
Ltac Debug
). Debugger output and prompts appear in the Messages panel (#13783, by Jim Fehrle and Emilio J. Gallego Arias).Added: Shift-return in the Find dialog now searches backwards (#13810, by slrnsc).
Standard library¶
Changed: Minor Changes to
Rpower
: Generalizesexp_ineq1
to hold for all non-zero numbers. Addsexp_ineq1_le
, which holds for all reals (but is a<=
instead of a<
). (#13582, by Avi Shinnar and Barry Trager, with help from Laurent Théry).Changed: set
n mod 0 = n
uniformly fornat
,N
,Z
,int63
,sint63
,int31
such thatm = (m / n) * n + (m mod n)
holds (also forn = 0
)Warning
code that relies on
n mod 0 = 0
will break; for compatibility with bothn mod 0 = n
andn mod 0 = 0
you can usen mod 0 = ltac:(match eval hnf in (1 mod 0) with |0 => exact 0 |_ => exact n end)
(#14086, by Andrej Dudenhefner with help of Guillaume Melquiond, Jason Gross, and Kazuhiko Sakaguchi).
Changed: The standard library now contains a more complete theory of equality on types of the form
exists x : A, P x
andexists2 x : A, P x & Q x
when we haveA : Prop
. To bring this theory more in line with the existing theory about sigma types,eq_ex_uncurried
,eq_ex2_uncurried
,eq_ex
,eq_ex2
,eq_ex_hprop
,eq_ex2_hprop
have been renamed intoeq_ex_intro_uncurried
,eq_ex_intro2_uncurried
,eq_ex_intro
,eq_ex_intro2
,eq_ex_intro_hprop
,eq_ex_intro2_hprop
respectively and the implicit status of these lemmas has changed slightly (#14174, by Jason Gross).Changed Moved 39 lemmas and notations about the rationals
Q
from the constructive reals private filetheories/Reals/Cauchy/QExtra.v
to appropriate files intheories/QArith
. The now public lemmas are mostly about compatibility of multiplication and power with relational operators and simple convenience lemmas e.g. for reduction ofQ
values. The following moved lemmas have been renamed:Q_factorDenom
toQmult_frac_l
,Q_reduce_fl
toQreduce_num_l
,Qle_neq
toQlt_leneq
,Qmult_lt_le_compat_nonneg
toQmult_le_lt_compat_pos
,Qpower_pos_lt
toQpower_0_lt
,Qpower_lt_1_increasing
toQpower_1_lt_pos
,Qpower_lt_1_increasing'
toQpower_1_lt
,Qpower_le_1_increasing
toQpower_1_le_pos
,Qpower_le_1_increasing'
toQpower_1_le
,Qzero_eq
toQreduce_zero
,Qpower_lt_compat
toQpower_lt_compat_l
,Qpower_le_compat
toQpower_le_compat_l
,Qpower_lt_compat_inv
toQpower_lt_compat_l_inv
,Qpower_le_compat_inv
toQpower_le_compat_l_inv
,Qpower_decomp'
toQpower_decomp_pos
andQarchimedeanExp2_Pos
toQarchimedean_power2_pos
. The following lemmas have been renamed and the sides of the equality swapped:Qinv_swap_pos
toQinv_pos
,Qinv_swap_neg
toQinv_neg
and. The following lemmas have been deleted:Q_factorNum_l
andQ_factorNum
. The lemmaQopp_lt_compat
has been moved fromtheories/QArith/Qround.v
totheories/QArith/QArith_base.v
. About 10 additional lemmas have been added for similar cases as the moved lemmas. Compatibility notations are not provided because QExtra is considered internal (excluded from the library documentation). (#14293, by Michael Soegtrop).Changed: Importing
ZArith
no longer has the side-effect of closingZ_scope
. (#14343, fixes #13307, by Ralf Jung).Removed:
IF_then_else
definition and correspondingIF P then Q else R
notation (#13871, by Yishuai Li).Removed: from
List.v
deprecated/unexpected dependenciesSetoid
,Le
,Gt
,Minus
,Lt
(#13986, by Andrej Dudenhefner).Deprecated: Unsigned primitive integers are now named
uint63
instead ofint63
. TheInt63
module is replaced byUint63
. The full list of changes is described in the PR. (#13895, by Ana Borges).Added:
leb
andltb
functions forascii
(#13080, by Yishuai Li).Added: Library for signed primitive integers, Sint63. The following operations were added to the kernel: division, remainder, comparison functions, and arithmetic shift right. Everything else works the same for signed and unsigned ints. (#13559, fixes #12109, by Ana Borges, Guillaume Melquiond and Pierre Roux).
Added: Lemmas about vectors related with
to_list
:length_to_list
,of_list_to_list_opp
,to_list_nil
,to_list_cons
,to_list_hd
,to_list_last
,to_list_const
,to_list_nth_order
,to_list_tl
,to_list_append
,to_list_rev_append_tail
,to_list_rev_append
,to_list_rev
,to_list_map
,to_list_fold_left
,to_list_fold_right
,to_list_Forall
,to_list_Exists
,to_list_In
,to_list_Forall2
(#13671, by Olivier Laurent).Added: Lemmas about
count_occ
:count_occ_app
,count_occ_elt_eq
,count_occ_elt_neq
,count_occ_bound
,count_occ_repeat_eq
,count_occ_repeat_neq
,count_occ_unique
,count_occ_repeat_excl
,count_occ_sgt
,Permutation_count_occ
(#13804, by Olivier Laurent with help of Jean-Christophe Léchenet).Added: Lemmas to
List
:Exists_map
,Exists_concat
,Exists_flat_map
,Forall_map
,Forall_concat
,Forall_flat_map
,nth_error_map
,nth_repeat
,nth_error_repeat
(#13955, by Andrej Dudenhefner, with help from Olivier Laurent).Added:
Cantor.v
containing the Cantor pairing function and its inverse.Cantor.to_nat : nat * nat -> nat
andCantor.of_nat : nat -> nat * nat
are the respective bijections betweennat * nat
andnat
. (#14008, by Andrej Dudenhefner).Added: Lemmas to
Q
:Qeq_from_parts
,Qden_cancel
,Qnum_cancel
,Qreduce_l
,Qreduce_r
,Qmult_inject_Z_l
,Qmult_inject_Z_r
QArith_base Reduction of rationals; establishing equality for Qden/Qnum separately. (#14087, by Karolin Varner).Added:
Coq.Structures.OrdersEx.String_as_OT
andCoq.Structures.OrdersEx.Ascii_as_OT
to make strings and ascii ordered types (using lexical order). (#14096, by Jason Gross).Added: Lemmas
app_eq_app
,Forall_nil_iff
,Forall_cons_iff
toList.v
(#14153, closes #1803, by Andrej Dudenhefner, with help from Olivier Laurent).Added:
Z
,positive
andN
constants can now be printed in hexadecimal by openinghex_Z_scope
,hex_positive_scope
, andhex_N_scope
respectively (#14263, by Jason Gross).Added: Absolute value function for Sint63. (#14384, by Ana Borges).
Added: Lemmas showing
firstn
andskipn
commute withmap
. (#14406, by Rudy Peterson).Fixed: Haskell extraction is now compatible with GHC versions >= 9.0. Some
#if
statements have been added to extractunsafeCoerce
to its new location in newer versions of GHC. (#14345, fixes #14256, by Jason Gross).
Infrastructure and dependencies¶
Changed: Coq's configure script now requires absolute paths for the
-prefix
option. (#12567, by Emilio Jesus Gallego Arias).Changed: The regular Coq package has been split in two: coq-core, with OCaml-based libraries and tools; and coq-stdlib, which contains the Gallina-based standard library. The package Coq now depends on both for compatiblity. (#12567, by Emilio Jesus Gallego Arias, review by Vincent Laporte, Guillaume Melquiond, Enrico Tassi, and Théo Zimmerman).
Changed: Coq's OCaml parts and tools [
coq-core
] are now built using Dune. The main user-facing change is that Dune >= 2.5 is now required to build Coq. This was a large and complex change. If you are packager you may find some minor differences if you were using a lot of custom optimizations. Note that, in particular, the configure option-datadir
is not customizable anymore, and-bindir
has been removed in favor of$prefix/bin
. Moreover, the install procedure will ignore-docdir
and-etcdir
, unless you patch the makefile and use Dune >= 2.9. We usually recommended using a recent Dune version, if possible. For developers and plugin authors, see the entry indev/doc/changes.md
. For packagers and users, seedev/doc/INSTALL.make.md
. (#13617, by Emilio Jesús Gallego Arias, Rudi Grinberg, and Théo Zimmerman; review and testing by Gaëtan Gilbert, Guillaume Melquiond, and Enrico Tassi)Changed: Undocumented variables
OLDROOT
andCOQPREFIXINSTALL
which added a prefix path tomake install
have been removed. Now,make install
does support the more standardDESTDIR
variable, akin to whatcoq_makefile
does. (#14258, by Emilio Jesus Gallego Arias).Added: Support OCaml 4.12 (#13885, by Emilio Jesus Gallego Arias, review by Gaëtan Gilbert and Théo Zimmerman).
Miscellaneous¶
Changed: The representation of micromega caches was slightly altered for efficiency purposes. As a consequence all stale caches must be cleaned up (#13405, by Pierre-Marie Pédrot).
Fixed: Fix the timeout facility on Unix to allow for nested timeouts. Previous behavior on nested timeouts was that an "inner" timeout would replace an "outer" timeout, so that the outer timeout would no longer fire. With the new behavior, Unix and Windows implementations should be (approximately) equivalent. (#13586, by Lasse Blaauwbroek).
Changes in 8.14.1¶
Kernel¶
Specification language, type inference¶
Fixed: Missing registration of universe constraints in
Module Type
elaboration (#14666, fixes #14505, by Hugo Herbelin).
Tactics¶
Commands and options¶
Fixed: anomaly with
Extraction Conservative Types
when extracting pattern-matching on singleton types (#14669, fixes #3527, by Hugo Herbelin).Fixed: a regular error instead of an anomaly when calling
Separate Extraction
in a module (#14670, fixes #10796, by Hugo Herbelin).
Version 8.13¶
Summary of changes¶
Coq version 8.13 integrates many usability improvements, as well as extensions of the core language. The main changes include:
Introduction of primitive persistent arrays in the core language, implemented using imperative persistent arrays.
Introduction of definitional proof irrelevance for the equality type defined in the SProp sort.
Cumulative record and inductive type declarations can now specify the variance of their universes.
Various bugfixes and uniformization of behavior with respect to the use of implicit arguments and the handling of existential variables in declarations, unification and tactics.
New warning for unused variables in catch-all match branches that match multiple distinct patterns.
New warning for
Hint
commands outside sections without a locality attribute, whose goal is to eventually remove the fragile default behavior of importing hints only when usingRequire
. The recommended fix is to declare hints asexport
, instead of the current defaultglobal
, meaning that they are imported throughRequire Import
only, notRequire
. See the following rationale and guidelines for details.General support for boolean attributes.
Many improvements to the handling of notations, including number notations, recursive notations and notations with bindings. A new algorithm chooses the most precise notation available to print an expression, which might introduce changes in printing behavior.
Tactic improvements in
lia
and itszify
preprocessing step, now supporting reasoning on boolean operators such asZ.leb
and supporting primitive integersInt63
.Typing flags can now be specified per-constant / inductive.
Improvements to the reference manual including updated syntax descriptions that match Coq's grammar in several chapters, and splitting parts of the tactics chapter to independent sections.
See the Changes in 8.13+beta1 section and following sections for the detailed list of changes, including potentially breaking changes marked with Changed.
Coq's documentation is available at https://coq.github.io/doc/v8.13/refman (reference manual), and https://coq.github.io/doc/v8.13/stdlib (documentation of the standard library). Developer documentation of the ML API is available at https://coq.github.io/doc/v8.13/api.
Maxime Dénès, Emilio Jesús Gallego Arias, Gaëtan Gilbert, Michael Soegtrop and Théo Zimmermann worked on maintaining and improving the continuous integration system and package building infrastructure.
Erik Martin-Dorel has maintained the Coq Docker images that are used in many Coq projects for continuous integration.
The OPAM repository for Coq packages has been maintained by Guillaume Claret, Karl Palmskog, Matthieu Sozeau and Enrico Tassi with contributions from many users. A list of packages is available at https://coq.inria.fr/opam/www/.
Our current 32 maintainers are Yves Bertot, Frédéric Besson, Tej Chajed, Cyril Cohen, Pierre Corbineau, Pierre Courtieu, Maxime Dénès, Jim Fehrle, Julien Forest, Emilio Jesús Gallego Arias, Gaëtan Gilbert, Georges Gonthier, Benjamin Grégoire, Jason Gross, Hugo Herbelin, Vincent Laporte, Olivier Laurent, Assia Mahboubi, Kenji Maillard, Guillaume Melquiond, Pierre-Marie Pédrot, Clément Pit-Claudel, Kazuhiko Sakaguchi, Vincent Semeria, Michael Soegtrop, Arnaud Spiwack, Matthieu Sozeau, Enrico Tassi, Laurent Théry, Anton Trunov, Li-yao Xia and Théo Zimmermann.
The 51 contributors to this version are Reynald Affeldt, Tanaka Akira, Frédéric Besson, Lasse Blaauwbroek, Clément Blaudeau, Martin Bodin, Ali Caglayan, Tej Chajed, Cyril Cohen, Julien Coolen, Matthew Dempsky, Maxime Dénès, Andres Erbsen, Jim Fehrle, Emilio Jesús Gallego Arias, Paolo G. Giarrusso, Attila Gáspár, Gaëtan Gilbert, Jason Gross, Benjamin Grégoire, Hugo Herbelin, Wolf Honore, Jasper Hugunin, Ignat Insarov, Ralf Jung, Fabian Kunze, Vincent Laporte, Olivier Laurent, Larry D. Lee Jr, Thomas Letan, Yishuai Li, James Lottes, Jean-Christophe Léchenet, Kenji Maillard, Erik Martin-Dorel, Yusuke Matsushita, Guillaume Melquiond, Carl Patenaude-Poulin, Clément Pit-Claudel, Pierre-Marie Pédrot, Pierre Roux, Kazuhiko Sakaguchi, Vincent Semeria, Michael Soegtrop, Matthieu Sozeau, Enrico Tassi, Anton Trunov, Edward Wang, Li-yao Xia, Beta Ziliani and Théo Zimmermann.
The Coq community at large helped improve the design of this new version via the GitHub issue and pull request system, the Coq development mailing list coqdev@inria.fr, the coq-club@inria.fr mailing list, the Discourse forum and the Coq Zulip chat.
Version 8.13's development spanned 5 months from the release of Coq 8.12.0. Enrico Tassi and Maxime Dénès are the release managers of Coq 8.13. This release is the result of 400 merged PRs, closing ~100 issues.
Changes in 8.13+beta1¶
Kernel¶
Added: Definitional UIP, only when
Definitional UIP
is enabled. This models definitional uniqueness of identity proofs for the equality type in SProp. It is deactivated by default as it can lead to non-termination in combination with impredicativity. Use of this flag is also printed byPrint Assumptions
. See documentation of the flag for details. (#10390, by Gaëtan Gilbert).Added: Built-in support for persistent arrays, which expose a functional interface but are implemented using an imperative data structure, for better performance. (#11604, by Maxime Dénès and Benjamin Grégoire, with help from Gaëtan Gilbert).
Primitive arrays are irrelevant in their single polymorphic universe (same as a polymorphic cumulative list inductive would be) (#13356, fixes #13354, by Gaëtan Gilbert).
Fixed: A loss of definitional equality for declarations obtained through
Include
when entering the scope of aModule
orModule Type
was causingSearch
not to see the included declarations (#12537, fixes #12525 and #12647, by Hugo Herbelin).Fixed: Fix an incompleteness in the typechecking of
match
for cumulative inductive types. This could result in breaking subject reduction. (#13501, fixes #13495, by Matthieu Sozeau).
Specification language, type inference¶
Changed: Boolean attributes are now specified using key/value pairs, that is to say
identattr= yesno?
. If the value is missing, the default isyes
. The old syntax is still supported, but produces thedeprecated-attribute-syntax
warning.Deprecated attributes are
universes(monomorphic)
,universes(notemplate)
anduniverses(noncumulative)
, which are respectively replaced byuniverses(polymorphic=no)
,universes(template=no)
anduniverses(cumulative=no)
. Attributesprogram
andcanonical
are also affected, with the syntaxidentattr(false)
being deprecated in favor ofidentattr=no
. (#13312, by Emilio Jesus Gallego Arias).Changed: Heuristics for universe minimization to
Set
: also use constraintsProp <= i
(#10331, by Gaëtan Gilbert with help from Maxime Dénès and Matthieu Sozeau, fixes #12414).Changed: The type given to
Instance
is no longer automatically generalized over unbound and generalizable variables. UseInstance : `{type}
instead ofInstance : type
to get the old behavior, or enable the compatibility flagInstance Generalized Output
. (#13188, fixes #6042, by Gaëtan Gilbert).Changed: Tweaked the algorithm giving default names to arguments. Should reduce the frequency that argument names get an unexpected suffix. Also makes
Mangle Names
not mess up argument names. (#12756, fixes #12001 and #6785, by Jasper Hugunin).Removed: Undocumented and experimental forward class hint feature
:>>
. Use:>
(seeof_type
) instead (#13106, by Pierre-Marie Pédrot).Added: Commands
Inductive
,Record
and synonyms now support syntaxInductive foo@{=i +j *k l}
to specify variance information for their universes (in Cumulative mode) (#12653, by Gaëtan Gilbert).Added: Warning on unused variables in pattern-matching branches of
match
serving as catch-all branches for at least two distinct patterns. (#12768, fixes #12762, by Hugo Herbelin).Added: Definition and (Co)Fixpoint now support the
using
attribute. It has the same effect asProof using
, which is only available in interactive mode. (#13183, by Enrico Tassi).Added: Typing flags can now be specified per-constant / inductive, this allows to fine-grain specify them from plugins or attributes. See Controlling Typing Flags for details on attribute syntax. (#12586, by Emilio Jesus Gallego Arias).
Added: Inference of return predicate of a
match
by inversion takes sort elimination constraints into account (#13290, grants #13278, by Hugo Herbelin).Fixed: Implicit arguments taken into account in defined fields of a record type declaration (#13166, fixes #13165, by Hugo Herbelin).
Fixed: Allow use of typeclass inference for the return predicate of a
match
(was deactivated in versions 8.10 to 8.12, #13217, fixes #13216, by Hugo Herbelin).Fixed: A case of unification raising an anomaly IllTypedInstance (#13376, fixes #13266, by Hugo Herbelin).
Fixed: Using
{wf ...}
in local fixpoints is an error, not an anomaly (#13383, fixes #11816, by Hugo Herbelin).Fixed: Issue when two expressions involving different projections and one is primitive need to be unified (#13386, fixes #9971, by Hugo Herbelin).
Fixed: A bug producing ill-typed instances of existential variables when let-ins interleaved with assumptions (#13387, fixes #12348, by Hugo Herbelin).
Notations¶
Changed: In notations (except in custom entries), the misleading
syntax_modifier
ident ident
(which accepted either an identifier or a_
) is deprecated and should be replaced byident name
. If the intent was really to only parse identifiers, this will eventually become possible, but only as of Coq 8.15. In custom entries, the meaning ofident ident
is silently changed from parsing identifiers or_
to parsing only identifiers without warning, but this presumably affects only rare, recent and relatively experimental code (#11841, fixes #9514, by Hugo Herbelin).Changed: Improved support for notations/abbreviations with mixed terms and patterns (such as the forcing modality) (#12099, by Hugo Herbelin).
Changed Rational and real constants are parsed differently. The exponent is now encoded separately from the fractional part using
Z.pow_pos
. This way, parsing large exponents can no longer blow up and constants are printed in a form closer to the one in which they were parsed (i.e.,102e-2
is reprinted as such and not1.02
). (#12218, by Pierre Roux).Changed: Scope information is propagated in indirect applications to a reference prefixed with
@
; this covers for instance the caser.(@p) t
where scope information fromp
is now taken into account for interpretingt
(#12685, by Hugo Herbelin).Changed: New model for
only parsing
andonly printing
notations with support for at most one parsing-and-printing or only-parsing notation per notation and scope, but an arbitrary number of only-printing notations (#12950, fixes #4738 and #9682 and part 2 of #12908, by Hugo Herbelin).Changed: Redeclaring a notation also reactivates its printing rule; in particular a second
Import
of the same module reactivates the printing rules declared in this module. In theory, this leads to changes in behavior for printing. However, this is mitigated in general by the adoption in #12986 of a priority given to notations which match a larger part of the term to print (#12984, fixes #7443 and #10824, by Hugo Herbelin).Changed: Use of notations for printing now gives preference to notations which match a larger part of the term to abbreviate (#12986, by Hugo Herbelin).
Removed OCaml parser and printer for real constants have been removed. Real constants are now handled with proven Coq code. (#12218, by Pierre Roux).
Deprecated
Numeral.v
is deprecated, please useNumber.v
instead. (#12218, by Pierre Roux).Deprecated:
Numeral Notation
, please useNumber Notation
instead (#12979, by Pierre Roux).Added:
Printing Float
flag to print primitive floats as hexadecimal instead of decimal values. This is included in thePrinting All
flag (#11986, by Pierre Roux).Added: Number Notation and String Notation commands now support parameterized inductive and non-inductive types (#12218, fixes #12035, by Pierre Roux, review by Jason Gross and Jim Fehrle for the reference manual).
Added: Added support for encoding notations of the form
x ⪯ y ⪯ .. ⪯ z ⪯ t
. This feature is considered experimental. (#12765, by Hugo Herbelin).Added: The
binder
entry ofNotation
can now be used in notations expecting a single (non-recursive) binder (#13265, by Hugo Herbelin, see section Notations and binders of the reference manual).Fixed: Issues in the presence of notations recursively referring to another applicative notations, such as missing scope propagation, or failure to use a notation for printing (#12960, fixes #9403 and #10803, by Hugo Herbelin).
Fixed: Capture the names of global references by binders in the presence of notations for binders (#12965, fixes #9569, by Hugo Herbelin).
Fixed: Preventing notations for constructors to involve binders (#13092, fixes #13078, by Hugo Herbelin).
Fixed: Notations understand universe names without getting confused by different imported modules between declaration and use locations (#13415, fixes #13303, by Gaëtan Gilbert).
Tactics¶
Changed: In
refine
, new existential variables unified with existing ones are no longer considered as fresh. The behavior ofsimple refine
no longer depends on the orientation of evar-evar unification problems, and new existential variables are always turned into (unshelved) goals. This can break compatibility in some cases (#7825, by Matthieu Sozeau, with help from Maxime Dénès, review by Pierre-Marie Pédrot and Enrico Tassi, fixes #4095 and #4413).Changed: Giving an empty list of occurrences after
in
in tactics is no longer permitted. Omitting thein
gives the same behavior (#13237, fixes #13235, by Hugo Herbelin).Removed:
at occs_nums
clauses in tactics such asunfold
no longer allow negative values. A "-" before the list (for set complement) is still supported. Ex: "at -1 -2" is no longer supported but "at -1 2" is. (#13403, by Jim Fehrle).Removed: A number of tactics that formerly accepted negative numbers as parameters now give syntax errors for negative values. These include {e}constructor, do, timeout, 9 {e}auto tactics and psatz*. (#13417, by Jim Fehrle).
Removed: The deprecated and undocumented
prolog
tactic was removed (#12399, by Pierre-Marie Pédrot).Removed:
info
tactic that was deprecated in 8.5. (#12423, by Jim Fehrle).Deprecated: Undocumented
eauto nat_or_var nat_or_var
syntax in favor of newbfs eauto
. Also deprecated 2-integer syntax fordebug eauto
andinfo_eauto
. (Usebfs eauto
with theInfo Eauto
orDebug Eauto
flags instead.) (#13381, by Jim Fehrle).Added:
lia
is extended to deal with boolean operators e.g.andb
orZ.leb
. (Aslia
gets more powerful, this may break proof scripts relying onlia
failure.) (#11906, by Frédéric Besson).Added:
apply … in
supports several hypotheses (#12246, by Hugo Herbelin; grants #9816).Added: The
zify
tactic can now be extended by redefining thezify_pre_hook
tactic. (#12552, by Kazuhiko Sakaguchi).Added: The
zify
tactic provides support for primitive integers (moduleZifyInt63
). (#12648, by Frédéric Besson).Fixed: Avoid exposing an internal name of the form
_tmp
when applying the_
introduction pattern which would break a dependency (#13337, fixes #13336, by Hugo Herbelin).Fixed: The case of tactics, such as
eapply
, producing existential variables under binders with an ill-formed instance (#13373, fixes #13363, by Hugo Herbelin).
Commands and options¶
Changed: Drop prefixes from grammar non-terminal names, e.g. "constr:global" -> "global", "Prim.name" -> "name". Visible in the output of
Print Grammar
andPrint Custom Grammar
. (#13096, by Jim Fehrle).Changed: When declaring arbitrary terms as hints, unsolved evars are not abstracted implicitly anymore and instead raise an error (#13139, by Pierre-Marie Pédrot).
Removed: In the
Extraction Language
command, removeOcaml
as a valid value. UseOCaml
instead. This was deprecated in Coq 8.8, #6261 (#13016, by Jim Fehrle).Deprecated: Hint locality currently defaults to
local
in a section andglobal
otherwise, but this will change in a future release. Hints added outside of sections without an explicit locality now generate a deprecation warning. We recommend usingexport
where possible (#13384, by Pierre-Marie Pédrot).Deprecated:
Grab Existential Variables
andExistential
commands (#12516, by Maxime Dénès).Added: The
export
locality can now be used for all Hint commands, includingHint Cut
,Hint Mode
,Hint Transparent
/Opaque
andRemove Hints
(#13388, by Pierre-Marie Pédrot).Added: Support for automatic insertion of coercions in
Search
patterns. Additionally, head patterns are now automatically interpreted as types (#13255, fixes #13244, by Hugo Herbelin).Added: The
Proof using
command can now be used without loading the Ltac plugin (-noinit
mode) (#13339, by Théo Zimmermann).Added: Clarify in the documentation that
Add ML Path
is not exported to compiled files (#13345, fixes #13344, by Hugo Herbelin).
Tools¶
Changed: Option
-native-compiler
of the configure script now impacts the default value of the-native-compiler
option of coqc. The-native-compiler
option of the configure script supports a newondemand
value, which becomes the default, thus preserving the previous default behavior. The stdlib is still precompiled when configuring with-native-compiler yes
. It is not precompiled otherwise. This an implementation of point 2 of CEP #48 (#13352, by Pierre Roux).Changed: Added the ability for coq_makefile to directly set the installation folders, through the
COQLIBINSTALL
andCOQDOCINSTALL
variables. See CoqMakefile.local. (#12389, by Martin Bodin, review of Enrico Tassi).Removed: The option
-I
of coqchk was removed (it was deprecated in Coq 8.8) (#12613, by Gaëtan Gilbert).Fixed:
coqchk
no longer reports names from inner modules of opaque modules as axioms (#12862, fixes #12845, by Jason Gross).
CoqIDE¶
Added: Support showing diffs for
Show Proof
in CoqIDE from theView
menu. See "Show Proof" differences. (#12874, by Jim Fehrle and Enrico Tassi)Added: Support for flag
Printing Goal Names
in View menu (#13145, by Hugo Herbelin).
Standard library¶
Changed: In the reals theory changed the epsilon in the definition of the modulus of convergence for CReal from 1/n (n in positive) to 2^z (z in Z) so that a precision coarser than one is possible. Also added an upper bound to CReal to enable more efficient computations. (#12186, by Michael Soegtrop).
Changed: Int63 notations now match up with the rest of the standard library:
a \% m
,m == n
,m < n
,m <= n
, andm ≤ n
have been replaced witha mod m
,m =? n
,m <? n
,m <=? n
, andm ≤? n
. The old notations are still available as deprecated notations. Additionally, there is now aCoq.Numbers.Cyclic.Int63.Int63.Int63Notations
module that users can import to get theInt63
notations without unqualifying the various primitives (#12479, fixes #12454, by Jason Gross).Changed: PrimFloat notations now match up with the rest of the standard library:
m == n
,m < n
, andm <= n
have been replaced withm =? n
,m <? n
, andm <=? n
. The old notations are still available as deprecated notations. Additionally, there is now aCoq.Floats.PrimFloat.PrimFloatNotations
module that users can import to get thePrimFloat
notations without unqualifying the various primitives (#12556, fixes #12454, by Jason Gross).Changed: the sort of cyclic numbers from Type to Set. For backward compatibility, a dynamic sort was defined in the 3 packages bignums, coqprime and color. See for example commit 6f62bda in bignums. (#12801, by Vincent Semeria).
Changed:
Require Import Coq.nsatz.NsatzTactic
now allows usingnsatz
withZ
andQ
without having to supply instances or usingRequire Import Coq.nsatz.Nsatz
, which transitively requires unneeded files declaring axioms used in the reals (#12861, fixes #12860, by Jason Gross).Deprecated:
prod_curry
andprod_uncurry
, in favor ofuncurry
andcurry
(#12716, by Yishuai Li).Added: New lemmas about
repeat
inList
andPermutation
:repeat_app
,repeat_eq_app
,repeat_eq_cons
,repeat_eq_elt
,Forall_eq_repeat
,Permutation_repeat
(#12799, by Olivier Laurent).Added: Extend some list lemmas to both directions:
app_inj_tail_iff
,app_inv_head_iff
,app_inv_tail_iff
. (#12094, fixes #12093, by Edward Wang).Added:
Decidable
instance for negation (#12420, by Yishuai Li).Fixed:
Coq.Program.Wf.Fix_F_inv
andCoq.Program.Wf.Fix_eq
are now axiom-free. They no longer assume proof irrelevance. (#13365, by Li-yao Xia).
Infrastructure and dependencies¶
Changed: When compiled with OCaml >= 4.10.0, Coq will use the new best-fit GC policy, which should provide some performance benefits. Coq's policy is optimized for speed, but could increase memory consumption in some cases. You are welcome to tune it using the
OCAMLRUNPARAM
variable and report back on good settings so we can improve the defaults. (#13040, fixes #11277, by Emilio Jesus Gallego Arias).Changed: Coq now uses the zarith library, based on GNU's gmp instead of
num
which is deprecated upstream. The custombigint
module is no longer provided. (#11742, #13007, by Emilio Jesus Gallego Arias and Vicent Laporte, with help from Frédéric Besson).
Changes in 8.13.0¶
Changes in 8.13.1¶
Changes in 8.13.2¶
Kernel¶
Fixed: Crash when using
vm_compute
on an irreduciblePArray.set
(#14005, fixes #13998, by Guillaume Melquiond).Fixed: Never store persistent arrays as VM / native structured values. This could be used to make vo marshalling crash, and probably breaking some other invariants of the kernel (#14007, fixes #14006, by Pierre-Marie Pédrot).
Version 8.12¶
Summary of changes¶
Coq version 8.12 integrates many usability improvements, in particular with respect to notations, scopes and implicit arguments, along with many bug fixes and major improvements to the reference manual. The main changes include:
New binder notation for non-maximal implicit arguments using
[ ]
allowing to set and see the implicit status of arguments immediately.New notation
Inductive I A | x : s := ...
to distinguish the uniform from the non-uniform parameters in inductive definitions.More robust and expressive treatment of implicit inductive parameters in inductive declarations.
Improvements in the treatment of implicit arguments and partially applied constants in notations, parsing of hexadecimal number notation and better handling of scopes and coercions for printing.
A correct and efficient coercion coherence checking algorithm, avoiding spurious or duplicate warnings.
An improved
Search
command which accepts complex queries. Note that this takes precedence over the now deprecated ssreflect search.Many additions and improvements of the standard library.
Improvements to the reference manual include a more logical organization of chapters along with updated syntax descriptions that match Coq's grammar in most but not all chapters.
Additionally, the omega
tactic is deprecated in this version of Coq,
and we recommend users to switch to lia
in new proof scripts.
See the Changes in 8.12+beta1 section and following sections for the detailed list of changes, including potentially breaking changes marked with Changed.
Coq's documentation is available at https://coq.github.io/doc/v8.12/refman (reference manual), and https://coq.github.io/doc/v8.12/stdlib (documentation of the standard library). Developer documentation of the ML API is available at https://coq.github.io/doc/v8.12/api.
Maxime Dénès, Emilio Jesús Gallego Arias, Gaëtan Gilbert, Michael Soegtrop and Théo Zimmermann worked on maintaining and improving the continuous integration system and package building infrastructure.
Erik Martin-Dorel has maintained the Coq Docker images that are used in many Coq projects for continuous integration.
The OPAM repository for Coq packages has been maintained by Guillaume Claret, Karl Palmskog, Matthieu Sozeau and Enrico Tassi with contributions from many users. A list of packages is available at https://coq.inria.fr/opam/www/.
Previously, most components of Coq had a single principal maintainer. This was changed in 8.12 (#11295) so that every component now has a team of maintainers, who are in charge of reviewing and merging incoming pull requests. This gave us a chance to significantly expand the pool of maintainters and provide faster feedback to contributors. Special thanks to all our maintainers!
Our current 31 maintainers are Yves Bertot, Frédéric Besson, Tej Chajed, Cyril Cohen, Pierre Corbineau, Pierre Courtieu, Maxime Dénès, Jim Fehrle, Julien Forest, Emilio Jesús Gallego Arias, Gaëtan Gilbert, Georges Gonthier, Benjamin Grégoire, Jason Gross, Hugo Herbelin, Vincent Laporte, Assia Mahboubi, Kenji Maillard, Guillaume Melquiond, Pierre-Marie Pédrot, Clément Pit-Claudel, Kazuhiko Sakaguchi, Vincent Semeria, Michael Soegtrop, Arnaud Spiwack, Matthieu Sozeau, Enrico Tassi, Laurent Théry, Anton Trunov, Li-yao Xia, Théo Zimmermann
The 59 contributors to this version are Abhishek Anand, Yves Bertot, Frédéric Besson, Lasse Blaauwbroek, Simon Boulier, Quentin Carbonneaux, Tej Chajed, Arthur Charguéraud, Cyril Cohen, Pierre Courtieu, Matthew Dempsky, Maxime Dénès, Andres Erbsen, Erika (@rrika), Nikita Eshkeev, Jim Fehrle, @formalize, Emilio Jesús Gallego Arias, Paolo G. Giarrusso, Gaëtan Gilbert, Jason Gross, Samuel Gruetter, Attila Gáspár, Hugo Herbelin, Jan-Oliver Kaiser, Robbert Krebbers, Vincent Laporte, Olivier Laurent, Xavier Leroy, Thomas Letan, Yishuai Li, Kenji Maillard, Erik Martin-Dorel, Guillaume Melquiond, Ike Mulder, Guillaume Munch-Maccagnoni, Antonio Nikishaev, Karl Palmskog, Pierre-Marie Pédrot, Clément Pit-Claudel, Ramkumar Ramachandra, Lars Rasmusson, Daniel de Rauglaudre, Talia Ringer, Pierre Roux, Kazuhiko Sakaguchi, Vincent Semeria, @scinart, Kartik Singhal, Michael Soegtrop, Matthieu Sozeau, Enrico Tassi, Laurent Théry, Ralf Treinen, Anton Trunov, Bernhard M. Wiedemann, Li-yao Xia, Nickolai Zeldovich and Théo Zimmermann.
Many power users helped to improve the design of this new version via the GitHub issue and pull request system, the Coq development mailing list coqdev@inria.fr, the coq-club@inria.fr mailing list, the Discourse forum and the new Coq Zulip chat (thanks to Cyril Cohen for organizing the move from Gitter).
Version 8.12's development spanned 6 months from the release of Coq 8.11.0. Emilio Jesus Gallego Arias and Théo Zimmermann are the release managers of Coq 8.12. This release is the result of ~500 PRs merged, closing ~100 issues.
Changes in 8.12+beta1¶
Specification language, type inference¶
Changed: The deprecation warning raised since Coq 8.10 when a trailing implicit is declared to be non-maximally inserted (with the command
Arguments
) has been turned into an error (#11368, by SimonBoulier).Changed: Typeclass resolution, accessible through
typeclasses eauto
, now suspends constraints according to their modes instead of failing. If a typeclass constraint does not match any of the declared modes for its class, the constraint is postponed, and the proof search continues on other goals. Proof search does a fixed point computation to try to solve them at a later stage of resolution. It does not fail if there remain only stuck constraints at the end of resolution. This makes typeclasses with declared modes more robust with respect to the order of resolution. (#10858, fixes #9058, by Matthieu Sozeau).Added: Warn when manual implicit arguments are used in unexpected positions of a term (e.g. in
Check id (forall {x}, x)
) or when an implicit argument name is shadowed (e.g. inCheck fun f : forall {x:nat} {x}, nat => f
) (#10202, by Hugo Herbelin).Added:
Arguments
now supports setting implicit an anonymous argument, as e.g. inArguments id {A} {_}
(#11098, by Hugo Herbelin, fixes #4696, #5173, #9098).Added: Syntax for non-maximal implicit arguments in definitions and terms using square brackets. The syntax is
[x : A]
,[x]
,`[A]
to be consistent with the commandArguments
(#11235, by Simon Boulier).Added:
Implicit Types
are now taken into account for printing. To inhibit it, unset thePrinting Use Implicit Types
flag (#11261, by Hugo Herbelin, granting #10366).Added: New syntax
Inductive
ident binder* | binder* := ...
to specify which parameters of an inductive type are uniform. See Parameterized inductive types (#11600, by Gaëtan Gilbert).Added: Warn when using
Fixpoint
orCoFixpoint
for definitions which are not recursive (#12121, by Hugo Herbelin).Fixed: More robust and expressive treatment of implicit inductive parameters in inductive declarations (#11579, by Maxime Dénès, Gaëtan Gilbert and Jasper Hugunin; fixes #7253 and #11585).
Fixed: Anomaly which could be raised when printing binders with implicit types (#12323, by Hugo Herbelin; fixes #12322).
Fixed: Case of an anomaly in trying to infer the return clause of an ill-typed
match
(#12422, fixes #12418, by Hugo Herbelin).
Notations¶
Changed: Notation scopes are now always inherited in notations binding a partially applied constant, including for notations binding an expression of the form
@qualid
. The latter was not the case beforehand (part of #11120).Changed: The printing algorithm now interleaves search for notations and removal of coercions (#11172, by Hugo Herbelin).
Changed: Nicer printing for decimal constants in R and Q. 1.5 is now printed 1.5 rather than 15e-1 (#11848, by Pierre Roux).
Removed: deprecated
compat
modifier ofNotation
andInfix
commands. Use thedeprecated
attribute instead (#11113, by Théo Zimmermann, with help from Jason Gross).Deprecated: Numeral Notation on
Decimal.uint
,Decimal.int
andDecimal.decimal
are replaced respectively by numeral notations onNumeral.uint
,Numeral.int
andNumeral.numeral
(#11948, by Pierre Roux).Added: Notations declared with the
where
clause in the declaration of inductive types, coinductive types, record fields, fixpoints and cofixpoints now support theonly parsing
modifier (#11602, by Hugo Herbelin).Added:
Printing Parentheses
flag to print parentheses even when implied by associativity or precedence (#11650, by Hugo Herbelin and Abhishek Anand).Added: Numeral notations now parse hexadecimal constants such as
0x2a
or0xb.2ap-2
. Parsers added fornat
,positive
,Z
,N
,Q
,R
, primitive integers and primitive floats (#11948, by Pierre Roux).Added: Abbreviations support arguments occurring both in term and binder position (#8808, by Hugo Herbelin).
Fixed: Different interpretations in different scopes of the same notation string can now be associated with different printing formats (#10832, by Hugo Herbelin, fixes #6092 and #7766).
Fixed: Parsing and printing consistently handle inheritance of implicit arguments in notations. With the exception of notations of the form
Notation string := @qualid
andNotation ident := @qualid
which inhibit implicit arguments, all notations binding a partially applied constant, as e.g. inNotation string := (qualid arg+)
, orNotation string := (@qualid arg+)
, orNotation ident := (qualid arg+)
, orNotation ident := (@qualid arg+)
, inherit the remaining implicit arguments (#11120, by Hugo Herbelin, fixing #4690 and #11091).Fixed: Notations in
only printing
mode do not uselessly reserve parsing keywords (#11590, by Hugo Herbelin, fixes #9741).Fixed: Numeral Notations now play better with multiple scopes for the same inductive type. Previously, when multiple numeral notations were defined for the same inductive, only the last one was considered for printing. Now, among the notations that are usable for printing and either have a scope delimiter or are open, the selection is made according to the order of open scopes, or according to the last defined notation if no appropriate scope is open (#12163, fixes #12159, by Pierre Roux, review by Hugo Herbelin and Jason Gross).
Tactics¶
Changed: The
rapply
tactic inCoq.Program.Tactics
now handles arbitrary numbers of underscores and takes in auconstr
. In rare cases where users were relying onrapply
inserting exactly 15 underscores and no more, due to the lemma having a completely unspecified codomain (and thus allowing for any number of underscores), the tactic will now loop instead (#10760, by Jason Gross).Changed: The
auto with zarith
tactic and variations (includingintuition
) may now calllia
instead ofomega
(when theOmega
module is loaded); more goals may be automatically solved, fewer section variables will be captured spuriously (#11018, by Vincent Laporte).Changed: The new
NativeCompute Timing
flag causes calls tonative_compute
(as well as kernel calls to the native compiler) to emit separate timing information about conversion to native code, compilation, execution, and reification. It replaces the timing information previously emitted when the-debug
command-line flag was set, and allows more fine-grained timing of the native compiler (#11025, by Jason Gross). Additionally, the timing information now uses real time rather than user time (fixes #11962, #11963, by Jason Gross)Changed: Improve the efficiency of
PreOmega.elim_let
using an iterator implemented in OCaml (#11370, by Frédéric Besson).Changed: Improve the efficiency of
zify
by rewritting the remaining Ltac code in OCaml (#11429, by Frédéric Besson).Changed: Backtrace information for tactics has been improved (#11755, by Emilio Jesus Gallego Arias).
Changed: The default tactic used by
firstorder
isauto with core
instead ofauto with *
; see Solvers for logic and equality for details; old behavior can be reset by using the-compat 8.12
command-line flag; to ease the migration of legacy code, the default solver can be set todebug auto with *
withSet Firstorder Solver debug auto with *
(#11760, by Vincent Laporte).Changed:
autounfold
no longer fails when theOpaque
command is used on constants in the hint databases (#11883, by Attila Gáspár).Changed: Tactics with qualified name of the form
Coq.Init.Notations
are now qualified with prefixCoq.Init.Ltac
; users of the-noinit
option should now importCoq.Init.Ltac
if they want to use Ltac (#12023, by Hugo Herbelin; minor source of incompatibilities).Changed: Tactic
subst
ident
now fails over a section variable which is indirectly dependent in the goal; the incompatibility can generally be fixed by first clearing the hypotheses causing an indirect dependency, as reported by the error message, or by usingrewrite
... in *
instead; similarly,subst
has no more effect on such variables (#12146, by Hugo Herbelin; fixes #10812 and #12139).Changed: The check that
unfold
arguments were indeed unfoldable has been moved to runtime (#12256, by Pierre-Marie Pédrot; fixes #5764, #5159, #4925 and #11727).Changed When the tactic
functional induction
c1 c2 ... cn
is used with no parenthesis aroundc1 c2 ... cn
,c1 c2 ... cn
is now read as one single applicative term. In particular implicit arguments should be omitted. Rare source of incompatibility (#12326, by Pierre Courtieu).Changed: When using
exists
oreexists
with multiple arguments, the evaluation of arguments and applications of constructors are now interleaved. This improves unification in some cases (#12366, fixes #12365, by Attila Gáspár).Removed: Undocumented
omega with
. Usinglia
is the recommended replacement, although the old semantics ofomega with *
can also be recovered withzify; omega
(#11288, by Emilio Jesus Gallego Arias).Removed: Deprecated syntax
_eqn
fordestruct
andremember
. Useeqn:
syntax instead (#11877, by Hugo Herbelin).Removed:
at
clauses can no longer be used withautounfold
. Since they had no effect, it is safe to remove them (#11883, by Attila Gáspár).Deprecated: The
omega
tactic is deprecated; uselia
from the Micromega plugin instead (#11976, by Vincent Laporte).Added: The
zify
tactic is now aware ofPos.pred_double
,Pos.pred_N
,Pos.of_nat
,Pos.add_carry
,Pos.pow
,Pos.square
,Z.pow
,Z.double
,Z.pred_double
,Z.succ_double
,Z.square
,Z.div2
, andZ.quot2
. Injections for internal definitions in moduleZifyBool
(isZero
andisLeZero
) are also added to help users to declare newzify
class instances using Micromega tactics (#10998, by Kazuhiko Sakaguchi).Added:
Show Lia Profile
prints some statistics aboutlia
calls (#11474, by Frédéric Besson).Added: Syntax
pose proof
(ident:=term)
as an alternative topose proof
term as ident
, following the model ofpose
(ident:=term)
(#11522, by Hugo Herbelin).Added: New tactical
with_strategy
which behaves like the commandStrategy
, with effects local to the given tactic (#12129, by Jason Gross).Added: The
zify
tactic is now aware ofNat.le
,Nat.lt
andNat.eq
(#12213, by Frédéric Besson; fixes #12210).Fixed:
zify
now handlesZ.pow_pos
by default. In Coq 8.11, this was the case only when loading moduleZifyPow
because this triggered a regression oflia
. The regression is now fixed, and the module kept only for compatibility (#11362, fixes #11191, by Frédéric Besson).Fixed: Efficiency regression of
lia
(#11474, fixes #11436, by Frédéric Besson).Fixed: The behavior of
autounfold
no longer depends on the names of terms and modules (#11883, fixes #7812, by Attila Gáspár).Fixed: Wrong type error in tactic
functional induction
(#12326, by Pierre Courtieu, fixes #11761, reported by Lasse Blaauwbroek).
Tactic language¶
Changed: The "reference" tactic generic argument now accepts arbitrary variables of the goal context (#12254, by Pierre-Marie Pédrot).
Added: An array library for Ltac2 (as compatible as possible with OCaml standard library) (#10343, by Michael Soegtrop).
Added: The Ltac2 rebinding command
Ltac2 Set
has been extended with the ability to give a name to the old value so as to be able to reuse it inside the new one (#11503, by Pierre-Marie Pédrot).Added: Ltac2 notations for
enough
andeenough
(#11740, by Michael Soegtrop).Added: New Ltac2 function
Fresh.Free.of_goal
to return the list of names of declarations of the current goal; new Ltac2 functionFresh.in_goal
to return a variable fresh in the current goal (#11882, by Hugo Herbelin).Added: Ltac2 notations for reductions in terms:
eval red_expr in term
(#11981, by Michael Soegtrop).Fixed: The
Ltac Profiling
machinery now correctly handles backtracking into multi-success tactics. The call-counts of some tactics are unfortunately inflated by 1, as some tactics are implicitly implemented astac + fail
, which has two entry-points rather than one (fixes #12196, #12197, by Jason Gross).
SSReflect¶
Changed: The
Search (ssreflect)
command that used to be available when loading thessreflect
plugin has been moved to a separate plugin that needs to be loaded separately:ssrsearch
(part of #8855, fixes #12253, by Théo Zimmermann).Deprecated:
Search (ssreflect)
(available throughRequire ssrsearch.
) in favor of theheadconcl:
clause ofSearch
(part of #8855, by Théo Zimmermann).
Flags, options and attributes¶
Changed: Legacy attributes can now be passed in any order (#11665, by Théo Zimmermann).
Removed:
Typeclasses Axioms Are Instances
flag, deprecated since 8.10. UseDeclare Instance
for axioms which should be instances (#11185, by Théo Zimmermann).Removed: Deprecated unsound compatibility
Template Check
flag that was introduced in 8.10 to help users gradually move their template polymorphic inductive type definitions outside sections (#11546, by Pierre-Marie Pédrot).Removed: Deprecated
Shrink Obligations
flag (#11828, by Emilio Jesus Gallego Arias).Removed: Unqualified
polymorphic
,monomorphic
,template
,notemplate
attributes (they were deprecated since Coq 8.10). Useuniverses(polymorphic)
,universes(monomorphic)
,universes(template)
anduniverses(notemplate)
instead (#11663, by Théo Zimmermann).Deprecated:
Hide Obligations
flag (#11828, by Emilio Jesus Gallego Arias).Added: Handle the
local
attribute inCanonical Structure
declarations (#11162, by Enrico Tassi).Added: New attributes supported when defining an inductive type
universes(cumulative)
,universes(noncumulative)
andprivate(matching)
, which correspond to legacy attributesCumulative
,NonCumulative
, and the previously undocumentedPrivate
(#11665, by Théo Zimmermann).Added: The Hint commands now accept the
export
locality as an attribute, allowing to make import-scoped hints (#11812, by Pierre-Marie Pédrot).Added:
Cumulative StrictProp
to control cumulativity of \(\SProp\) (#12034, by Gaëtan Gilbert).
Commands¶
Changed: The
Coercion
command has been improved to check the coherence of the inheritance graph. It checks whether a circular inheritance path ofC >-> C
is convertible with the identity function or not, then report it as an ambiguous path if it is not. The new mechanism does not report ambiguous paths that are redundant with others. For example, checking the ambiguity of[f; g]
and[f'; g]
is redundant with that of[f]
and[f']
thus will not be reported (#11258, by Kazuhiko Sakaguchi).Changed: Several commands (
Search
,About
, ...) now print the implicit arguments in brackets when printing types (#11795, by Simon Boulier).Changed: The warning when using
Require
inside a section moved from thedeprecated
category to thefragile
category, because there is no plan to remove the functionality at this time (#11972, by Gaëtan Gilbert).Changed:
Redirect
now obeys thePrinting Width
andPrinting Depth
options (#12358, by Emilio Jesus Gallego Arias).Removed: Recursive OCaml loadpaths are not supported anymore; the command
Add Rec ML Path
has been removed;Add ML Path
is now the preferred one. We have also dropped support for the non-qualified version of theAdd LoadPath
command, that is to say, theAdd LoadPath dir
version; now, you must always specify a prefix now usingAdd Loadpath dir as Prefix
(#11618, by Emilio Jesus Gallego Arias).Removed: undocumented
Chapter
command. UseSection
instead (#11746, by Théo Zimmermann).Removed:
SearchAbout
command that was deprecated since 8.5. UseSearch
instead (#11944, by Jim Fehrle).Deprecated: Declaration of arbitrary terms as hints. Global references are now preferred (#7791, by Pierre-Marie Pédrot).
Deprecated:
SearchHead
in favor of the newheadconcl:
clause ofSearch
(part of #8855, by Théo Zimmermann).Added:
Print Canonical Projections
can now take constants as arguments and prints only the unification rules that involve or are synthesized from the given constants (#10747, by Kazuhiko Sakaguchi).Added: A section variable introduced with
Let
can be declared as aCanonical Structure
(#11164, by Enrico Tassi).Added: Support for universe bindings and universe contrainsts in
Let
definitions (#11534, by Théo Zimmermann).Added: Support for new clauses
hyp:
,headhyp:
,concl:
,headconcl:
,head:
andis:
inSearch
. Support for complex search queries combining disjunctions, conjunctions and negations (#8855, by Hugo Herbelin, with ideas from Cyril Cohen and help from Théo Zimmermann).Fixed: A printing bug in the presence of elimination principles with local definitions (#12295, by Hugo Herbelin; fixes #12233).
Fixed: Anomalies with
Show Proof
(#12296, by Hugo Herbelin; fixes #12234).
Tools¶
Changed: Internal options and behavior of
coqdep
.coqdep
no longer works as a replacement forocamldep
, thus.ml
files are not supported as input. Also, several deprecated options have been removed:-w
,-D
,-mldep
,-prefix
,-slash
, and-dumpbox
. Passing-boot
tocoqdep
will not load any path by default now,-R/-Q
should be used instead (#11523 and #11589, by Emilio Jesus Gallego Arias).Changed: The order in which the require flags
-ri
,-re
,-rfrom
, etc. and the option flags-set
,-unset
are given now matters. In particular, it is now possible to interleave the loading of plugins and the setting of options by choosing the right order for these flags. The load flags-l
and-lv
are still processed afterward for now (#11851 and #12097, by Lasse Blaauwbroek).Changed: The
cleanall
target of a makefile generated bycoq_makefile
now erases.lia.cache
and.nia.cache
(#12006, by Olivier Laurent).Changed: The output of
make TIMED=1
(and therefore the timing targets such asprint-pretty-timed
andprint-pretty-timed-diff
) now displays the full name of the output file being built, rather than the stem of the rule (which was usually the filename without the extension, but in general could be anything for user-defined rules involving%
) (#12126, by Jason Gross).Changed: When passing
TIMED=1
tomake
with either Coq's own makefile or acoq_makefile
-made makefile, timing information is now printed for OCaml files as well (#12211, by Jason Gross).Changed: The pretty-timed scripts and targets now print a newline at the end of their tables, rather than creating text with no trailing newline (#12368, by Jason Gross).
Removed: The
-load-ml-source
and-load-ml-object
command-line options have been removed; their use was very limited, you can achieve the same adding additional object files in the linking step or using a plugin (#11409, by Emilio Jesus Gallego Arias).Removed: The confusingly-named
-require
command-line option, which was deprecated since 8.11. Use the equivalent-require-import
/-ri
options instead (#12005, by Théo Zimmermann).Deprecated:
-cumulative-sprop
command-line flag in favor of the newCumulative StrictProp
flag (#12034, by Gaëtan Gilbert).Added: A new documentation environment
details
to make certain portion of a Coq document foldable. See Hiding / Showing parts of the source (#10592, by Thomas Letan).Added: The
make-both-single-timing-files.py
script now accepts a--fuzz=N
parameter on the command line which determines how many characters two lines may be offset in the "before" and "after" timing logs while still being considered the same line. When invoking this script via theprint-pretty-single-time-diff
target in aMakefile
made bycoq_makefile
, you can set this argument by passingTIMING_FUZZ=N
tomake
(#11302, by Jason Gross).Added: The
make-one-time-file.py
andmake-both-time-files.py
scripts now accept a--real
parameter on the command line to print real times rather than user times in the tables. Themake-both-single-timing-files.py
script accepts a--user
parameter to use user times. When invoking these scripts via theprint-pretty-timed
orprint-pretty-timed-diff
orprint-pretty-single-time-diff
targets in aMakefile
made bycoq_makefile
, you can set this argument by passingTIMING_REAL=1
(to pass--real
) orTIMING_REAL=0
(to pass--user
) tomake
(#11302, by Jason Gross).Added: Coq's build system now supports both
TIMING_FUZZ
,TIMING_SORT_BY
, andTIMING_REAL
just like aMakefile
made bycoq_makefile
(#11302, by Jason Gross).Added: The
make-one-time-file.py
andmake-both-time-files.py
scripts now include peak memory usage information in the tables (can be turned off by the--no-include-mem
command-line parameter), and a--sort-by-mem
parameter to sort the tables by memory rather than time. When invoking these scripts via theprint-pretty-timed
orprint-pretty-timed-diff
targets in aMakefile
made bycoq_makefile
, you can set this argument by passingTIMING_INCLUDE_MEM=0
(to pass--no-include-mem
) andTIMING_SORT_BY_MEM=1
(to pass--sort-by-mem
) tomake
(#11606, by Jason Gross).Added: Coq's build system now supports both
TIMING_INCLUDE_MEM
andTIMING_SORT_BY_MEM
just like aMakefile
made bycoq_makefile
(#11606, by Jason Gross).Added: New
coqc
/coqtop
option-boot
that will not bind theCoq
library prefix by default (#11617, by Emilio Jesus Gallego Arias).Added: Definitions in coqdoc link to themselves, giving access in html to their own url (#12026, by Hugo Herbelin; granting #7093).
Added: Hyperlinks on bound variables in coqdoc (#12033, by Hugo Herbelin; it incidentally fixes #7697).
Added: Highlighting of link targets in coqdoc (#12091, by Hugo Herbelin).
Fixed: The various timing targets for Coq's standard library now correctly display and label the "before" and "after" columns, rather than mixing them up (#11302 fixes #11301, by Jason Gross).
Fixed: The sorting order of the timing script
make-both-time-files.py
and the targetprint-pretty-timed-diff
is now deterministic even when the sorting order isabsolute
ordiff
; previously the relative ordering of two files with identical times was non-deterministic (#11606, by Jason Gross).Fixed: Fields of a record tuple now link in coqdoc to their definition (#12027, fixes #3415, by Hugo Herbelin).
Fixed:
coqdoc
now reports the location of a mismatched opening[[
instead of throwing an uninformative exception (#12037, fixes #9670, by Xia Li-yao).Fixed: coqchk incorrectly reporting names from opaque modules as axioms (#12076, by Pierre Roux; fixes #5030).
Fixed: coq_makefile-generated
Makefile
spretty-timed-diff
target no longer raises Python exceptions in the rare corner case where the log of times contains no files (#12388, fixes #12387, by Jason Gross).
Standard library¶
Changed: Notations
[|term|]
and[||term||]
for morphisms from 63-bit integers toZ
andzn2z int
have been removed in favor ofφ(term)
andΦ(term)
respectively. These notations were breaking Ltac parsing (#11686, by Maxime Dénès).Changed: The names of
Sorted_sort
andLocallySorted_sort
inCoq.Sorting.MergeSort
have been swapped to appropriately reflect their meanings (#11885, by Lysxia).Changed: Notations
<=?
and<?
fromCoq.Structures.Orders
andCoq.Sorting.Mergesort.NatOrder
are now at level 70 rather than 35, so as to be compatible with the notations defined everywhere else in the standard library. This may require re-parenthesizing some expressions. These notations were breaking the ability to import modules from the standard library that were otherwise compatible (fixes #11890, #11891, by Jason Gross).Changed: The level of
≡
inCoq.Numbers.Cyclic.Int63.Int63
is now 70, no associativity, in line with=
. Note that this is a minor incompatibility with developments that declare their own≡
notation and importInt63
(fixes #11905, #11909, by Jason Gross).Changed: No longer re-export
ListNotations
fromProgram
(Program.Syntax
) (#11992, by Antonio Nikishaev).Changed: It is now possible to import the
nsatz
machinery without transitively depending on the axioms of the real numbers nor of classical logic by loadingCoq.nsatz.NsatzTactic
rather thanCoq.nsatz.Nsatz
. Note that some constants have changed kernel names, living inCoq.nsatz.NsatzTactic
rather thanCoq.nsatz.Nsatz
; this might cause minor incompatibilities that can be fixed by actually runningImport Nsatz
rather than relying on absolute names (#12073, by Jason Gross; fixes #5445).Changed: new lemma
NoDup_incl_NoDup
inList.v
to remove useless hypothesisNoDup l'
inSorting.Permutation.NoDup_Permutation_bis
(#12120, by Olivier Laurent).Changed:
Fixpoints
of the standard library without a recursive call turned into ordinaryDefinitions
(#12121, by Hugo Herbelin; fixes #11903).Deprecated:
Bool.leb
in favor ofBool.le
. The definition ofBool.le
is made local to avoid conflicts withNat.le
. As a consequence, previous calls toleb
based on importingBool
should now be qualified intoBool.le
even ifBool
is imported (#12162, by Olivier Laurent).Added: Theorem
bezout_comm
for natural numbers (#11127, by Daniel de Rauglaudre).Added
rew dependent
notations for the dependent version ofrew
inCoq.Init.Logic.EqNotations
to improve the display and parsing ofmatch
statements onLogic.eq
(#11240, by Jason Gross).Added: Lemmas about lists:
properties of
In
:in_elt
,in_elt_inv
properties of
nth
:app_nth2_plus
,nth_middle
,nth_ext
properties of
last
:last_last
,removelast_last
properties of
remove
:remove_cons
,remove_app
,notin_remove
,in_remove
,in_in_remove
,remove_remove_comm
,remove_remove_eq
,remove_length_le
,remove_length_lt
properties of
concat
:in_concat
,remove_concat
properties of
map
andflat_map
:map_last
,map_eq_cons
,map_eq_app
,flat_map_app
,flat_map_ext
,nth_nth_nth_map
properties of
incl
:incl_nil_l
,incl_l_nil
,incl_cons_inv
,incl_app_app
,incl_app_inv
,remove_incl
,incl_map
,incl_filter
,incl_Forall_in_iff
properties of
NoDup
andnodup
:NoDup_rev
,NoDup_filter
,nodup_incl
properties of
Exists
andForall
:Exists_nth
,Exists_app
,Exists_rev
,Exists_fold_right
,incl_Exists
,Forall_nth
,Forall_app
,Forall_elt
,Forall_rev
,Forall_fold_right
,incl_Forall
,map_ext_Forall
,Exists_or
,Exists_or_inv
,Forall_and
,Forall_and_inv
,exists_Forall
,Forall_image
,concat_nil_Forall
,in_flat_map_Exists
,notin_flat_map_Forall
properties of
repeat
:repeat_cons
,repeat_to_concat
definitions and properties of
list_sum
andlist_max
:list_sum_app
,list_max_app
,list_max_le
,list_max_lt
misc:
elt_eq_unit
,last_length
,rev_eq_app
,removelast_firstn_len
,cons_seq
,seq_S
Added: Well-founded induction principles for
nat
:lt_wf_rect1
,lt_wf_rect
,gt_wf_rect
,lt_wf_double_rect
(#11335, by Olivier Laurent).Added:
remove'
andcount_occ'
over lists, alternatives toremove
andcount_occ
based onfilter
(#11350, by Yishuai Li).Added: Facts about
N.iter
andPos.iter
:N.iter_swap_gen
,N.iter_swap
,N.iter_succ
,N.iter_succ_r
,N.iter_add
,N.iter_ind
,N.iter_invariant
;Pos.iter_succ_r
,Pos.iter_ind
.
(#11880, by Lysxia).
Added: Facts about
Permutation
:structure:
Permutation_refl'
,Permutation_morph_transp
compatibilities:
Permutation_app_rot
,Permutation_app_swap_app
,Permutation_app_middle
,Permutation_middle2
,Permutation_elt
,Permutation_Forall
,Permutation_Exists
,Permutation_Forall2
,Permutation_flat_map
,Permutation_list_sum
,Permutation_list_max
inversions:
Permutation_app_inv_m
,Permutation_vs_elt_inv
,Permutation_vs_cons_inv
,Permutation_vs_cons_cons_inv
,Permutation_map_inv
,Permutation_image
,Permutation_elt_map_inv
length-preserving definition by means of transpositions
Permutation_transp
with associated properties:Permutation_transp_sym
,Permutation_transp_equiv
,Permutation_transp_cons
,Permutation_Permutation_transp
,Permutation_ind_transp
(#11946, by Olivier Laurent).
Added: Notations for sigma types:
{ x & P & Q }
,{ ' pat & P }
,{ ' pat & P & Q }
(#11957, by Olivier Laurent).Added: Order relations
lt
andcompare
added inBool.Bool
. Order properties forbool
added inBool.BoolOrder
as well as two modulesBool_as_OT
andBool_as_DT
inStructures.OrdersEx
(#12008, by Olivier Laurent).Added: Properties of some operations on vectors:
nth_order
:nth_order_hd
,nth_order_tl
,nth_order_ext
replace
:nth_order_replace_eq
,nth_order_replace_neq
,replace_id
,replace_replace_eq
,replace_replace_neq
map
:map_id
,map_map
,map_ext_in
,map_ext
Forall
andForall2
:Forall_impl
,Forall_forall
,Forall_nth_order
,Forall2_nth_order
(#12014, by Olivier Laurent).
Added: Lemmas
orb_negb_l
,andb_negb_l
,implb_true_iff
,implb_false_iff
,implb_true_r
,implb_false_r
,implb_true_l
,implb_false_l
,implb_same
,implb_contrapositive
,implb_negb
,implb_curry
,implb_andb_distrib_r
,implb_orb_distrib_r
,implb_orb_distrib_l
in libraryBool
(#12018, by Hugo Herbelin).Added: Definition and properties of cyclic permutations / circular shifts:
CPermutation
(#12031, by Olivier Laurent).Added:
Structures.OrderedTypeEx.Ascii_as_OT
(#12044, by formalize.eth (formalize@protonmail.com)).Fixed: Rewrote
Structures.OrderedTypeEx.String_as_OT.compare
to avoid huge proof terms (#12044, by formalize.eth (formalize@protonmail.com); fixes #12015).
Reals library¶
Changed: Cleanup of names in the Reals theory: replaced
tan_is_inj
withtan_inj
and replacedatan_right_inv
withtan_atan
- compatibility notations are provided. Moved various auxiliary lemmas fromRatan.v
to more appropriate places (#9803, by Laurent Théry and Michael Soegtrop).Changed: Replace
CRzero
andCRone
byCR_of_Q 0
andCR_of_Q 1
inConstructiveReals
. Use implicit arguments forConstructiveReals
. MoveConstructiveReals
into new directoryAbstract
. Remove imports of implementations inside thoseAbstract
files. Move implementation by means of Cauchy sequences in new directoryCauchy
. Split filesConstructiveMinMax
andConstructivePower
.Warning
The constructive reals modules are marked as experimental.
Removed: Type
RList
has been removed. All uses have been replaced bylist R
. Functions fromRList
namedIn
,Rlength
,cons_Rlist
,app_Rlist
have also been removed as they are essentially the same asIn
,length
,app
, andmap
fromList
, modulo the following changes:RList.In x (RList.cons a l)
used to be convertible to(x = a) \/ RList.In x l
, butList.In x (a :: l)
is convertible to(a = x) \/ List.In l
. The equality is reversed.app_Rlist
andList.map
take arguments in different order.
(#11404, by Yves Bertot).
Added: inverse trigonometric functions
asin
andacos
with lemmas for the derivatives, bounds and special values of these functions; an extensive set of identities between trigonometric functions and their inverse functions; lemmas for the injectivity of sine and cosine; lemmas on the derivative of the inverse of decreasing functions and on the derivative of horizontally mirrored functions; various generic auxiliary lemmas and definitions forRsqr
,sqrt
,posreal
and others (#9803, by Laurent Théry and Michael Soegtrop).
Extraction¶
Added: Support for better extraction of strings in OCaml and Haskell:
ExtOcamlNativeString
provides bindings from the CoqString
type to the OCamlstring
type, and string literals can be extracted to literals, both in OCaml and Haskell (#10486, by Xavier Leroy, with help from Maxime Dénès, review by Hugo Herbelin).Fixed: In Haskell extraction with
ExtrHaskellString
, equality comparisons on strings and characters are now guaranteed to be uniquely well-typed, even in very polymorphic contexts underunsafeCoerce
; this is achieved by adding type annotations to the extracted code, and by makingExtrHaskellString
exportExtrHaskellBasic
(#12263, by Jason Gross, fixes #12257 and #12258).
Reference manual¶
Changed: The reference manual has been restructured to get a more logical organization. In the new version, there are fewer top-level chapters, and, in the HTML format, chapters are split into smaller pages. This is still a work in progress and further restructuring is expected in the next versions of Coq (CEP#43, implemented in #11601, #11871, #11914, #12148, #12172, #12239 and #12330, effort inspired by Matthieu Sozeau, led by Théo Zimmermann, with help and reviews of Jim Fehrle, Clément Pit-Claudel and others).
Changed: Most of the grammar is now presented using the notation mechanism that has been used to present commands and tactics since Coq 8.8 and which is documented in Syntax conventions (#11183, #11314, #11423, #11705, #11718, #11720, #11961 and #12103, by Jim Fehrle, reviewed by Théo Zimmermann).
Added: A glossary of terms and an index of attributes (#11869, #12150 and #12224, by Jim Fehrle and Théo Zimmermann, reviewed by Clément Pit-Claudel)
Added: A selector that allows switching between versions of the reference manual (#12286, by Clément Pit-Claudel).
Fixed: Most of the documented syntax has been thoroughly updated to make it accurate and easily understood. This was done using a semi-automated
doc_grammar
tool introduced for this purpose and through significant revisions to the text (#9884, #10614, #11314, #11423, #11705, #11718, #11720 #11797, #11913, #11958, #11960, #11961 and #12103, by Jim Fehrle, reviewed by Théo Zimmermann and Jason Gross).
Infrastructure and dependencies¶
Changed: Minimal versions of dependencies for building the reference manual: now requires Sphinx >= 2.3.1 & < 3.0.0, sphinx_rtd_theme 0.4.3+ and sphinxcontrib-bibtex 0.4.2+.
Warning
The reference manual is known not to build properly with Sphinx 3.
(#12224, by Jim Fehrle and Théo Zimmermann).
Removed: Python 2 is no longer required in any part of the codebase (#11245, by Emilio Jesus Gallego Arias).
Changes in 8.12.0¶
Notations
Added: Simultaneous definition of terms and notations now support custom entries. Fixes #11121. (#12523, by Maxime Dénès).
Fixed: Printing bug with notations for n-ary applications used with applied references. (#12683, fixes #12682, by Hugo Herbelin).
Tactics
Fixed:
typeclasses eauto
(and discriminated hint bases) now correctly classify local variables as being unfoldable (#12572, fixes #12571, by Pierre-Marie Pédrot).
Tactic language
Fixed: Excluding occurrences was causing an anomaly in tactics (e.g.,
pattern _ at L
whereL
is-2
). (#12541, fixes #12228, by Pierre Roux).Fixed: Parsing of multi-parameters Ltac2 types (#12594, fixes #12595, by Pierre-Marie Pédrot).
SSReflect
Fixed: Do not store the full environment inside ssr ast_closure_term (#12708, fixes #12707, by Pierre-Marie Pédrot).
Commands and options
Fixed: Properly report the mismatched magic number of vo files (#12677, fixes #12513, by Pierre-Marie Pédrot).
Changed: Arbitrary hints have been undeprecated, and their definition now triggers a standard warning instead (#12678, fixes #11970, by Pierre-Marie Pédrot).
CoqIDE
Fixed: CoqIDE no longer exits when trying to open a file whose name is not a valid identifier (#12562, fixes #10988, by Vincent Laporte).
Infrastructure and dependencies
Changes in 8.12.1¶
Kernel
Fixed: Incompleteness of conversion checking on problems involving η-expansion and cumulative universe polymorphic inductive types (#12738, fixes #7015, by Gaëtan Gilbert).
Fixed: Polymorphic side-effects inside monomorphic definitions were incorrectly handled as not inlined. This allowed deriving an inconsistency (#13331, fixes #13330, by Pierre-Marie Pédrot).
Notations
Fixed: Undetected collision between a lonely notation and a notation in scope at printing time (#12946, fixes the first part of #12908, by Hugo Herbelin).
Fixed: Printing of notations in custom entries with variables not mentioning an explicit level (#13026, fixes #12775 and #13018, by Hugo Herbelin).
Tactics
Added:
replace
andinversion
support registration of acore.identity
-like equality inType
, such as HoTT'spath
(#12847, partially fixes #12846, by Hugo Herbelin).Fixed: Anomaly with
injection
involving artificial dependencies disappearing by reduction (#12816, fixes #12787, by Hugo Herbelin).
Tactic language
Fixed: Miscellaneous issues with locating tactic errors (#13247, fixes #12773 and #12992, by Hugo Herbelin).
SSReflect
Fixed: Regression in error reporting after
case
. A generic error message "Could not fill dependent hole in apply" was reported for any error followingcase
orelim
(#12857, fixes #12837, by Enrico Tassi).
Commands and options
Fixed: Failures of
Search
in the presence of primitive projections (#13301, fixes #13298, by Hugo Herbelin).Fixed:
Search
supports filtering on parts of identifiers which are not proper identifiers themselves, such as"1"
(#13351, fixes #13349, by Hugo Herbelin).
Tools
Fixed: Special symbols now escaped in the index produced by coqdoc, avoiding collision with the syntax of the output format (#12754, fixes #12752, by Hugo Herbelin).
Fixed: The
details
environment added in the 8.12 release can now be used as advertised in the reference manual (#12772, by Thomas Letan).Fixed: Targets such as
print-pretty-timed
incoq_makefile
-madeMakefile
s no longer error in rare cases where--output-sync
is not passed to make and the timing output gets interleaved in just the wrong way (#13063, fixes #13062, by Jason Gross).
CoqIDE
Fixed: View menu "Display parentheses" (#12794 and #13067, fixes #12793, by Jean-Christophe Léchenet and Hugo Herbelin).
Infrastructure and dependencies
Version 8.11¶
Summary of changes¶
The main changes brought by Coq version 8.11 are:
Ltac2, a new tactic language for writing more robust larger scale tactics, with built-in support for datatypes and the multi-goal tactic monad.
Primitive floats are integrated in terms and follow the binary64 format of the IEEE 754 standard, as specified in the
Coq.Float.Floats
library.Cleanups of the section mechanism, delayed proofs and further restrictions of template polymorphism to fix soundness issues related to universes.
New unsafe flags to disable locally guard, positivity and universe checking. Reliance on these flags is always printed by
Print Assumptions
.Fixed bugs of
Export
andImport
that can have a significant impact on user developments (common source of incompatibility!).New interactive development method based on
vos
interface files, allowing to work on a file without recompiling the proof parts of their dependencies.New
Arguments
annotation for bidirectional type inference configuration for reference (e.g. constants, inductive) applications.New refine attribute for
Instance
can be used instead of the removedRefine Instance Mode
.Generalization of the
under
andover
tactics of SSReflect to arbitrary relations.Revision of the
Coq.Reals
library, its axiomatisation and instances of the constructive and classical real numbers.
Additionally, while the omega
tactic is not yet deprecated in
this version of Coq, it should soon be the case and we already
recommend users to switch to lia
in new proof scripts.
The dev/doc/critical-bugs
file documents the known critical bugs
of Coq and affected releases. See the Changes in 8.11+beta1
section and following sections for the detailed list of changes,
including potentially breaking changes marked with Changed.
Coq's documentation is available at https://coq.github.io/doc/v8.11/api (documentation of the ML API), https://coq.github.io/doc/v8.11/refman (reference manual), and https://coq.github.io/doc/v8.11/stdlib (documentation of the standard library).
Maxime Dénès, Emilio Jesús Gallego Arias, Gaëtan Gilbert, Michael Soegtrop and Théo Zimmermann worked on maintaining and improving the continuous integration system and package building infrastructure.
The OPAM repository for Coq packages has been maintained by Guillaume Claret, Karl Palmskog, Matthieu Sozeau and Enrico Tassi with contributions from many users. A list of packages is available at https://coq.inria.fr/opam/www/.
The 61 contributors to this version are Michael D. Adams, Guillaume Allais, Helge Bahmann, Langston Barrett, Guillaume Bertholon, Frédéric Besson, Simon Boulier, Michele Caci, Tej Chajed, Arthur Charguéraud, Cyril Cohen, Frédéric Dabrowski, Arthur Azevedo de Amorim, Maxime Dénès, Nikita Eshkeev, Jim Fehrle, Emilio Jesús Gallego Arias, Paolo G. Giarrusso, Gaëtan Gilbert, Georges Gonthier, Jason Gross, Samuel Gruetter, Armaël Guéneau, Hugo Herbelin, Florent Hivert, Jasper Hugunin, Shachar Itzhaky, Jan-Oliver Kaiser, Robbert Krebbers, Vincent Laporte, Olivier Laurent, Samuel Lelièvre, Nicholas Lewycky, Yishuai Li, Jose Fernando Lopez Fernandez, Andreas Lynge, Kenji Maillard, Erik Martin-Dorel, Guillaume Melquiond, Alexandre Moine, Oliver Nash, Wojciech Nawrocki, Antonio Nikishaev, Pierre-Marie Pédrot, Clément Pit-Claudel, Lars Rasmusson, Robert Rand, Talia Ringer, JP Rodi, Pierre Roux, Kazuhiko Sakaguchi, Vincent Semeria, Michael Soegtrop, Matthieu Sozeau, spanjel, Claude Stolze, Enrico Tassi, Laurent Théry, James R. Wilcox, Xia Li-yao, Théo Zimmermann
Many power users helped to improve the design of the new features via the issue and pull request system, the Coq development mailing list, the coq-club@inria.fr mailing list or the Discourse forum. It would be impossible to mention exhaustively the names of everybody who to some extent influenced the development.
Version 8.11 is the sixth release of Coq developed on a time-based development cycle. Its development spanned 3 months from the release of Coq 8.10. Pierre-Marie Pédrot is the release manager and maintainer of this release, assisted by Matthieu Sozeau. This release is the result of 2000+ commits and 300+ PRs merged, closing 75+ issues.
Changes in 8.11+beta1¶
Kernel
Added: A built-in support of floating-point arithmetic, allowing one to devise efficient reflection tactics involving numerical computation. Primitive floats are added in the language of terms, following the binary64 format of the IEEE 754 standard, and the related operations are implemented for the different reduction engines of Coq by using the corresponding processor operators in rounding-to-nearest-even. The properties of these operators are axiomatized in the theory
Coq.Floats.FloatAxioms
which is part of the libraryCoq.Floats.Floats
. See Section Primitive Floats (#9867, closes #8276, by Guillaume Bertholon, Erik Martin-Dorel, Pierre Roux).Changed: Internal definitions generated by
abstract
-like tactics are now inlined inside universeQed
-terminated polymorphic definitions, similarly to what happens for their monomorphic counterparts, (#10439, by Pierre-Marie Pédrot).Fixed: Section data is now part of the kernel. Solves a soundness issue in interactive mode where global monomorphic universe constraints would be dropped when forcing a delayed opaque proof inside a polymorphic section. Also relaxes the nesting criterion for sections, as polymorphic sections can now appear inside a monomorphic one (#10664, by Pierre-Marie Pédrot).
Changed: Using
SProp
is now allowed by default, without needing to pass-allow-sprop
or useAllow StrictProp
(#10811, by Gaëtan Gilbert).
Specification language, type inference
Added: Annotation in
Arguments
for bidirectionality hints: it is now possible to tell type inference to use type information from the context once then
first arguments of an application are known. The syntax is:Arguments foo x y & z
. See Bidirectionality hints (#10049, by Maxime Dénès with help from Enrico Tassi).Added: Record fields can be annotated to prevent them from being used as canonical projections; see Canonical Structures for details (#10076, by Vincent Laporte).
Changed: Require parentheses around nested disjunctive patterns, so that pattern and term syntax are consistent; match branch patterns no longer require parentheses for notation at level 100 or more.
Warning
Incompatibilities
In
match p with (_, (0|1)) => ...
parentheses may no longer be omitted around0|1
.Notation
(p | q)
now potentially clashes with core pattern syntax, and should be avoided.-w disj-pattern-notation
flags suchNotation
.
See Extended pattern matching for details (#10167, by Georges Gonthier).
Changed:
Function
always opens a proof when used with ameasure
orwf
annotation, see Advanced recursive functions for the updated documentation (#10215, by Enrico Tassi).Changed: The legacy command
Add Morphism
always opens a proof and cannot be used inside a module type. In order to declare a module type parameter that happens to be a morphism, useDeclare Morphism
. See Deprecated syntax and backward incompatibilities for the updated documentation (#10215, by Enrico Tassi).Changed: The universe polymorphism setting now applies from the opening of a section. In particular, it is not possible anymore to mix polymorphic and monomorphic definitions in a section when there are no variables nor universe constraints defined in this section. This makes the behavior consistent with the documentation. (#10441, by Pierre-Marie Pédrot)
Added: The
Section
command now accepts the "universes" attribute. In addition to setting the section universe polymorphism, it also locally sets the universe polymorphic option inside the section. (#10441, by Pierre-Marie Pédrot)Fixed:
Program Fixpoint
now usesex
andsig
to make telescopes involvingProp
types (#10758, by Gaëtan Gilbert, fixing #10757 reported by Xavier Leroy).Changed: Output of the
Print
andAbout
commands. Arguments meta-data is now displayed as the correspondingArguments
command instead of the human-targeted prose used in previous Coq versions. (#10985, by Gaëtan Gilbert).Added:
refine
attribute forInstance
, a more predictable version of the oldRefine Instance Mode
which unconditionally opens a proof (#10996, by Gaëtan Gilbert).Changed: The unsupported attribute error is now an error-by-default warning, meaning it can be disabled (#10997, by Gaëtan Gilbert).
Fixed: Bugs sometimes preventing to define valid (co)fixpoints with implicit arguments in the presence of local definitions, see #3282 (#11132, by Hugo Herbelin).
Example
The following features an implicit argument after a local definition. It was wrongly rejected.
- Definition f := fix f (o := true) {n : nat} m {struct m} := match m with 0 => 0 | S m' => f (n:=n+1) m' end.
- f is defined
Notations
Added: Numeral Notations now support sorts in the input to printing functions (e.g., numeral notations can be defined for terms containing things like
@cons Set nat nil
). (#9883, by Jason Gross).Added: The
Notation
andInfix
commands now support thedeprecated
attribute (#10180, by Maxime Dénès).Deprecated: The former
compat
annotation for notations is deprecated, and its semantics changed. It is now made equivalent to using adeprecated
attribute, and is no longer connected with the-compat
command-line flag (#10180, by Maxime Dénès).Changed: A simplification of parsing rules could cause a slight change of parsing precedences for the very rare users who defined notations with
constr
at level strictly between 100 and 200 and used these notations on the right-hand side of a cast operator (:
,<:
,<<:
) (#10963, by Théo Zimmermann, simplification initially noticed by Jim Fehrle).
Tactics
Added: Syntax
injection term as [= intropattern+ ]
as an alternative toinjection term as simple_intropattern+
using the standard injection intropattern syntax (#9288, by Hugo Herbelin).Changed: Reimplementation of the
zify
tactic. The tactic is more efficient and copes with dependent hypotheses. It can also be extended by redefining the tacticzify_post_hook
. (#9856, fixes #8898, #7886, #9848 and #5155, by Frédéric Besson).Changed: The goal selector tactical
only
now checks that the goal range it is given is valid instead of ignoring goals out of the focus range (#10318, by Gaëtan Gilbert).Added: Flags
Lia Cache
,Nia Cache
andNra Cache
. (#10765, by Frédéric Besson, see #10772 for use case).Added: The
zify
tactic is now aware ofZ.to_N
. (#10774, grants #9162, by Kazuhiko Sakaguchi).Changed: The
assert_succeeds
andassert_fails
tactics now only run their tactic argument once, even if it has multiple successes. This prevents blow-up and looping from using multisuccess tactics withassert_succeeds
. (#10966 fixes #10965, by Jason Gross).Fixed: The
assert_succeeds
andassert_fails
tactics now behave correctly when their tactic fully solves the goal. (#10966 fixes #9114, by Jason Gross).
Tactic language
Added: Ltac2, a new version of the tactic language Ltac, that doesn't preserve backward compatibility, has been integrated in the main Coq distribution. It is still experimental, but we already recommend users of advanced Ltac to start using it and report bugs or request enhancements. See its documentation in the dedicated chapter (#10002, plugin authored by Pierre-Marie Pédrot, with contributions by various users, integration by Maxime Dénès, help on integrating / improving the documentation by Théo Zimmermann and Jim Fehrle).
Added: Ltac2 tactic notations with “constr” arguments can specify the notation scope for these arguments; see Notations for details (#10289, by Vincent Laporte).
Changed: White spaces are forbidden in the
&ident
syntax for ltac2 references that are described in Built-in quotations (#10324, fixes #10088, authored by Pierre-Marie Pédrot).
SSReflect
Added: Generalize tactics
under
andover
for any registered relation. More precisely, assume the given context lemma has typeforall f1 f2, .. -> (forall i, R1 (f1 i) (f2 i)) -> R2 f1 f2
. The first step performed byunder
(since Coq 8.10) amounts to calling the tacticrewrite
, which itself relies onsetoid_rewrite
if need be. So this step was already compatible with a double implication or setoid equality for the conclusion head symbolR2
. But a further step consists in tagging the generated subgoalR1 (f1 i) (?f2 i)
to protect it from unwanted evar instantiation, and getUnder_rel _ R1 (f1 i) (?f2 i)
that is displayed as'Under[ f1 i ]
. In Coq 8.10, this second (convenience) step was only performed whenR1
was Leibniz'eq
oriff
. Now, it is also performed for any relationR1
which has aRewriteRelation
instance (aRelationClasses.Reflexive
instance being also needed soover
can discharge the'Under[ _ ]
goal by instantiating the hidden evar.) This feature generalizing support for setoid-like relations is enabled as soon as we do bothRequire Import ssreflect.
andRequire Setoid.
Finally, a rewrite ruleUnderE
has been added if one wants to "unprotect" the evar, and instantiate it manually with another rule than reflexivity (i.e., without using theover
tactic nor theover
rewrite rule). See also Section Rewriting under binders (#10022, by Erik Martin-Dorel, with suggestions and review by Enrico Tassi and Cyril Cohen).Added: A
void
notation for the standard library empty type (Empty_set
) (#10932, by Arthur Azevedo de Amorim).Added: Lemma
inj_compr
tossr.ssrfun
(#11136, by Cyril Cohen).
Commands and options
Removed: Deprecated flag
Refine Instance Mode
(#9530, fixes #3632, #3890 and #4638 by Maxime Dénès, review by Gaëtan Gilbert).Changed:
Fail
does not catch critical errors (including "stack overflow") anymore (#10173, by Gaëtan Gilbert).Removed: Undocumented
Instance : !type
syntax (#10185, by Gaëtan Gilbert).Removed: Deprecated
Show Script
command (#10277, by Gaëtan Gilbert).Added: Unsafe commands to enable/disable guard checking, positivity checking and universes checking (providing a local
-type-in-type
). See Controlling Typing Flags (#10291 by Simon Boulier).Fixed: Two bugs in
Export
. This can have an impact on the behavior of theImport
command on libraries.Import A
whenA
importsB
which exportsC
was importingC
, whereasImport
is not transitive. Also, afterImport A B
, the import ofB
was sometimes incomplete (#10476, by Maxime Dénès).Warning
This is a common source of incompatibilities in projects migrating to Coq 8.11.
Changed: Output generated by
Printing Dependent Evars Line
flag used by the Prooftree tool in Proof General. (#10489, closes #4504, #10399 and #10400, by Jim Fehrle).Added: Optionally highlight the differences between successive proof steps in the
Show Proof
command. Experimental; only available in coqtop and Proof General for now, may be supported in other IDEs in the future. (#10494, by Jim Fehrle).Removed: Legacy commands
AddPath
,AddRecPath
, andDelPath
which were undocumented, broken variants ofAdd LoadPath
,Add Rec LoadPath
, andRemove LoadPath
(#11187, by Maxime Dénès and Théo Zimmermann).
Tools
Added:
coqc
now provides the ability to generate compiled interfaces. Usecoqc -vos foo.v
to skip all opaque proofs during the compilation offoo.v
, and output a file calledfoo.vos
. This feature is experimental. It enables working on a Coq file without the need to first compile the proofs contained in its dependencies (#8642 by Arthur Charguéraud, review by Maxime Dénès and Emilio Gallego).Added: Command-line options
-require-import
,-require-export
,-require-import-from
and-require-export-from
, as well as their shorthand,-ri
,-re
,-refrom
and -rifrom
. Deprecate confusing command line option-require
(#10245 by Hugo Herbelin, review by Emilio Gallego).Changed: Renamed
VDFILE
from.coqdeps.d
to.<CoqMakefile>.d
in thecoq_makefile
utility, where<CoqMakefile>
is the name of the output file given by the-o
option. In this way two generated makefiles can coexist in the same directory. (#10947, by Kazuhiko Sakaguchi).Fixed:
coq_makefile
now supports environment variableCOQBIN
with no ending/
character (#11068, by Gaëtan Gilbert).
Standard library
Changed: Moved the
auto
hints of theOrderedType
module into a newordered_type
database (#9772, by Vincent Laporte).Removed: Deprecated modules
Coq.ZArith.Zlogarithm
andCoq.ZArith.Zsqrt_compat
(#9811, by Vincent Laporte).Added: Module
Reals.Cauchy.ConstructiveCauchyReals
defines constructive real numbers by Cauchy sequences of rational numbers (#10445, by Vincent Semeria, with the help and review of Guillaume Melquiond and Bas Spitters). This module is not meant to be imported directly, please importReals.Abstract.ConstructiveReals
instead.Added: New module
Reals.ClassicalDedekindReals
defines Dedekind real numbers as boolean-valued functions along with 3 logical axioms: limited principle of omniscience, excluded middle of negations, and functional extensionality. The exposed typeR
in moduleReals.Rdefinitions
now corresponds to these Dedekind reals, hidden behind an opaque module, which significantly reduces the number of axioms needed (seeReals.Rdefinitions
andReals.Raxioms
), while preserving backward compatibility. Classical Dedekind reals are a quotient of constructive reals, which allows to transport many constructive proofs to the classical case (#10827, by Vincent Semeria, based on discussions with Guillaume Melquiond, Bas Spitters and Hugo Herbelin, code review by Hugo Herbelin).Added: New lemmas on
combine
,filter
,nodup
,nth
, andnth_error
functions on lists (#10651, and #10731, by Oliver Nash).Changed: The lemma
filter_app
was moved to theList
module (#10651, by Oliver Nash).Added: Standard equivalence between weak excluded-middle and the classical instance of De Morgan's law, in module
ClassicalFacts
(#10895, by Hugo Herbelin).
Infrastructure and dependencies
Changed: Coq now officially supports OCaml 4.08. See
INSTALL
file for details (#10471, by Emilio Jesús Gallego Arias).
Changes in 8.11.0¶
Kernel
Changed: the native compilation (
native_compute
) now creates a directory to contain temporary files instead of putting them in the root of the system temporary directory (#11081, by Gaëtan Gilbert).Fixed: #11360. Broken section closing when a template polymorphic inductive type depends on a section variable through its parameters (#11361, by Gaëtan Gilbert).
Fixed: The type of
Set+1
would be computed to be itself, leading to a proof of False (#11422, by Gaëtan Gilbert).
Specification language, type inference
Changed: Heuristics for universe minimization to
Set
: only minimize flexible universes (#10657, by Gaëtan Gilbert with help from Maxime Dénès and Matthieu Sozeau).Fixed: A dependency was missing when looking for default clauses in the algorithm for printing pattern matching clauses (#11233, by Hugo Herbelin, fixing #11231, reported by Barry Jay).
Notations
Fixed:
Print Visibility
was failing in the presence of only-printing notations (#11276, by Hugo Herbelin, fixing #10750).Fixed: Recursive notations with custom entries were incorrectly parsing
constr
instead of custom grammars (#11311 by Maxime Dénès, fixes #9532, #9490).
Tactics
Changed: The tactics
eapply
,refine
and variants no longer allow shelved goals to be solved by typeclass resolution (#10762, by Matthieu Sozeau).Fixed: The optional string argument to
time
is now properly quoted underPrint Ltac
(#11203, fixes #10971, by Jason Gross)Fixed: Efficiency regression of
lia
introduced in 8.10 by PR #9725 (#11263, fixes #11063, and #11242, and #11270, by Frédéric Besson).Deprecated: The undocumented
omega with
tactic variant has been deprecated. Usinglia
is the recommended replacement, though the old semantics ofomega with *
can be recovered withzify; omega
(#11337, by Emilio Jesus Gallego Arias).Fixed For compatibility reasons, in 8.11,
zify
does not supportZ.pow_pos
by default. It can be enabled by explicitly loading the moduleZifyPow
(#11430 by Frédéric Besson fixes #11191).
Tactic language
Fixed: Syntax of tactic
cofix ... with ...
was broken since Coq 8.10 (#11241, by Hugo Herbelin).
Commands and options
Deprecated: The
-load-ml-source
and-load-ml-object
command line options have been deprecated; their use was very limited, you can achieve the same by adding object files in the linking step or by using a plugin (#11428, by Emilio Jesus Gallego Arias).
Tools
Fixed:
coqtop --version
was broken when called in the middle of an installation process (#11255, by Hugo Herbelin, fixing #11254).Deprecated: The
-quick
command is renamed to-vio
, for consistency with the new-vos
and-vok
flags. Usage of-quick
is now deprecated (#11280, by Arthur Charguéraud).Fixed:
coq_makefile
does not break when using theCAMLPKGS
variable together with an unpacked (mllib
) plugin (#11357, by Gaëtan Gilbert).Fixed:
coqdoc
with option-g
(Gallina only) now correctly prints commands with attributes (#11394, fixes #11353, by Karl Palmskog).
CoqIDE
Changed: CoqIDE now uses the GtkSourceView native implementation of the autocomplete mechanism (#11400, by Pierre-Marie Pédrot).
Standard library
Removed: Export of module
RList
inRanalysis
andRanalysis_reg
. ModuleRList
is still there but must be imported explicitly where required (#11396, by Michael Soegtrop).
Infrastructure and dependencies
Added: Build date can now be overridden by setting the
SOURCE_DATE_EPOCH
environment variable (#11227, by Bernhard M. Wiedemann).
Changes in 8.11.1¶
Kernel
Fixed: Allow more inductive types in
Unset Positivity Checking
mode (#11811, by SimonBoulier).
Notations
Fixed: Bugs in dealing with precedences of notations in custom entries (#11530, by Hugo Herbelin, fixing in particular #9517, #9519, #9521, #11331).
Added: In primitive floats, print a warning when parsing a decimal value that is not exactly a binary64 floating-point number. For instance, parsing 0.1 will print a warning whereas parsing 0.5 won't. (#11859, by Pierre Roux).
CoqIDE
Infrastructure and dependencies
Added: Bump official OCaml support and CI testing to 4.10.0 (#11131, #11123, #11102, by Emilio Jesus Gallego Arias, Jacques-Henri Jourdan, Guillaume Melquiond, and Guillaume Munch-Maccagnoni).
Miscellaneous
Fixed:
Extraction Implicit
on the constructor of a record was leading to an anomaly (#11329, by Hugo Herbelin, fixes #11114).
Changes in 8.11.2¶
Kernel
Fixed: Using
Require
inside a section caused an anomaly when closing the section. (#11972, by Gaëtan Gilbert, fixing #11783, reported by Attila Boros).
Tactics
Fixed: Anomaly with induction schemes whose conclusion is not normalized (#12116, by Hugo Herbelin; fixes #12045)
Fixed: Loss of location of some tactic errors (#12223, by Hugo Herbelin; fixes #12152 and #12255).
Commands and options
Changed: Ignore -native-compiler option when built without native compute support. (#12070, by Pierre Roux).
CoqIDE
Changed: CoqIDE now uses native window frames by default on Windows. The GTK window frames can be restored by setting the
GTK_CSD
environment variable to1
(#12060, fixes #11080, by Attila Gáspár).Fixed: New patch presumably fixing the random Coq 8.11 segfault issue with CoqIDE completion (#12068, by Hugo Herbelin, presumably fixing #11943).
Fixed: Highlighting style consistently applied to all three buffers of CoqIDE (#12106, by Hugo Herbelin; fixes #11506).
Version 8.10¶
Summary of changes¶
Coq version 8.10 contains two major new features: support for a native fixed-precision integer type and a new sort \(\SProp\) of strict propositions. It is also the result of refinements and stabilization of previous features, deprecations or removals of deprecated features, cleanups of the internals of the system and API, and many documentation improvements. This release includes many user-visible changes, including deprecations that are documented in the next subsection, and new features that are documented in the reference manual. Here are the most important user-visible changes:
Kernel:
A notion of primitive object was added to the calculus. Its first instance is primitive cyclic unsigned integers, axiomatized in module
UInt63
. See Section Primitive Integers. TheCoq.Numbers.Cyclic.Int31
library is deprecated (#6914, by Maxime Dénès, Benjamin Grégoire and Vincent Laporte, with help and reviews from many others).The \(\SProp\) sort of definitionally proof-irrelevant propositions was introduced. \(\SProp\) allows to mark proof terms as irrelevant for conversion, and is treated like \(\Prop\) during extraction. It is enabled using the
-allow-sprop
command-line flag or theAllow StrictProp
flag. See Chapter SProp (proof irrelevant propositions) (#8817, by Gaëtan Gilbert).The unfolding heuristic in termination checking was made more complete, allowing more constants to be unfolded to discover valid recursive calls. Performance regression may occur in Fixpoint declarations without an explicit
{struct}
annotation, since guessing the decreasing argument can now be more expensive (#9602, by Enrico Tassi).
Universes:
Added Subgraph variant to
Print Universes
. Try for instancePrint Universes Subgraph(sigT2.u1 sigT_of_sigT2.u1 projT3_eq.u1).
(#8451, by Gaëtan Gilbert).Added private universes for opaque polymorphic constants, see the documentation for the
Private Polymorphic Universes
flag, and unset it to get the previous behavior (#8850, by Gaëtan Gilbert).
Notations:
New command
String Notation
to register string syntax for custom inductive types (#8965, by Jason Gross).Experimental: Number Notations now parse decimal constants such as
1.02e+01
or10.2
. Parsers added forQ
andR
. In the rare case when such numeral notations were used in a development along withQ
orR
, they may have to be removed or disambiguated through explicit scope annotations (#8764, by Pierre Roux).
Ltac backtraces can be turned on using the
Ltac Backtrace
flag, which is off by default (#9142, fixes #7769 and #7385, by Pierre-Marie Pédrot).The tactics
lia
,nia
,lra
,nra
are now using a novel Simplex-based proof engine. In case of regression, unsetSimplex
to get the venerable Fourier-based engine (#8457, by Fréderic Besson).SSReflect:
New intro patterns:
temporary introduction:
=> +
block introduction:
=> [^ prefix ] [^~ suffix ]
fast introduction:
=> >
tactics as views:
=> /ltac:mytac
replace hypothesis:
=> {}H
See Section Introduction in the context (#6705, by Enrico Tassi, with help from Maxime Dénès, ideas coming from various users).
New tactic
under
to rewrite under binders, given an extensionality lemma:It can take occurrence switches, contextual patterns, and intro patterns:
under {2}[in RHS]eq_big => [i|i ?]
(#9651, by Erik Martin-Dorel and Enrico Tassi).
Combined Scheme
now works when inductive schemes are generated in sort \(\Type\). It used to be limited to sortProp
(#7634, by Théo Winterhalter).A new registration mechanism for reference from ML code to Coq constructs has been added (#186, by Emilio Jesús Gallego Arias, Maxime Dénès and Vincent Laporte).
CoqIDE:
CoqIDE now depends on gtk+3 and lablgtk3 instead of gtk+2 and lablgtk2. The INSTALL file available in the Coq sources has been updated to list the new dependencies (#9279, by Hugo Herbelin, with help from Jacques Garrigue, Emilio Jesús Gallego Arias, Michael Sogetrop and Vincent Laporte).
Smart input for Unicode characters. For example, typing
\alpha
thenShift+Space
will insert the greek letter alpha. A larger number of default bindings are provided, following the latex naming convention. Bindings can be customized, either globally, or on a per-project basis. See Section Bindings for input of Unicode symbols for details (#8560, by Arthur Charguéraud).
Infrastructure and dependencies:
Coq 8.10 requires OCaml >= 4.05.0, bumped from 4.02.3 See the
INSTALL
file for more information on dependencies (#7522, by Emilio Jesús Gallego Arías).Coq 8.10 doesn't need Camlp5 to build anymore. It now includes a fork of the core parsing library that Coq uses, which is a small subset of the whole Camlp5 distribution. In particular, this subset doesn't depend on the OCaml AST, allowing easier compilation and testing on experimental OCaml versions. Coq also ships a new parser
coqpp
that plugin authors must switch to (#7902, #7979, #8161, #8667, and #8945, by Pierre-Marie Pédrot and Emilio Jesús Gallego Arias).The Coq developers would like to thank Daniel de Rauglaudre for many years of continued support.
Coq now supports building with Dune, in addition to the traditional Makefile which is scheduled for deprecation (#6857, by Emilio Jesús Gallego Arias, with help from Rudi Grinberg).
Experimental support for building Coq projects has been integrated in Dune at the same time, providing an improved experience for plugin developers. We thank the Dune team for their work supporting Coq.
Version 8.10 also comes with a bunch of smaller-scale changes and improvements regarding the different components of the system, including many additions to the standard library (see the next subsection for details).
On the implementation side, the dev/doc/changes.md
file documents
the numerous changes to the implementation and improvements of
interfaces. The file provides guidelines on porting a plugin to the new
version and a plugin development tutorial originally made by Yves Bertot
is now in doc/plugin_tutorial
. The dev/doc/critical-bugs
file
documents the known critical bugs of Coq and affected releases.
The efficiency of the whole system has seen improvements thanks to contributions from Gaëtan Gilbert, Pierre-Marie Pédrot, and Maxime Dénès.
Maxime Dénès, Emilio Jesús Gallego Arias, Gaëtan Gilbert, Michael Soegtrop, Théo Zimmermann worked on maintaining and improving the continuous integration system and package building infrastructure. Coq is now continuously tested against the OCaml trunk, in addition to the oldest supported and latest OCaml releases.
Coq's documentation for the development branch is now deployed
continuously at https://coq.github.io/doc/master/api (documentation of
the ML API), https://coq.github.io/doc/master/refman (reference
manual), and https://coq.github.io/doc/master/stdlib (documentation of
the standard library). Similar links exist for the v8.10
branch.
The OPAM repository for Coq packages has been maintained by Guillaume Melquiond, Matthieu Sozeau, Enrico Tassi (who migrated it to opam 2) with contributions from many users. A list of packages is available at https://coq.inria.fr/opam/www/.
The 61 contributors to this version are Tanaka Akira, Benjamin Barenblat, Yves Bertot, Frédéric Besson, Lasse Blaauwbroek, Martin Bodin, Joachim Breitner, Tej Chajed, Frédéric Chapoton, Arthur Charguéraud, Cyril Cohen, Lukasz Czajka, David A. Dalrymple, Christian Doczkal, Maxime Dénès, Andres Erbsen, Jim Fehrle, Emilio Jesus Gallego Arias, Gaëtan Gilbert, Matěj Grabovský, Simon Gregersen, Jason Gross, Samuel Gruetter, Hugo Herbelin, Jasper Hugunin, Mirai Ikebuchi, Chantal Keller, Matej Košík, Sam Pablo Kuper, Vincent Laporte, Olivier Laurent, Larry Darryl Lee Jr, Nick Lewycky, Yao Li, Yishuai Li, Assia Mahboubi, Simon Marechal, Erik Martin-Dorel, Thierry Martinez, Guillaume Melquiond, Kayla Ngan, Karl Palmskog, Pierre-Marie Pédrot, Clément Pit-Claudel, Pierre Roux, Kazuhiko Sakaguchi, Ryan Scott, Vincent Semeria, Gan Shen, Michael Soegtrop, Matthieu Sozeau, Enrico Tassi, Laurent Théry, Kamil Trzciński, whitequark, Théo Winterhalter, Xia Li-yao, Beta Ziliani and Théo Zimmermann.
Many power users helped to improve the design of the new features via the issue and pull request system, the Coq development mailing list, the coq-club@inria.fr mailing list or the new Discourse forum. It would be impossible to mention exhaustively the names of everybody who to some extent influenced the development.
Version 8.10 is the fifth release of Coq developed on a time-based development cycle. Its development spanned 6 months from the release of Coq 8.9. Vincent Laporte is the release manager and maintainer of this release. This release is the result of ~2500 commits and ~650 PRs merged, closing 150+ issues.
Other changes in 8.10+beta1¶
Command-line tools and options:
The use of
coqtop
as a compiler has been deprecated, in favor ofcoqc
. Consequently option-compile
will stop to be accepted in the next release.coqtop
is now reserved to interactive use (#9095, by Emilio Jesús Gallego Arias).New option
-topfile filename
, which will set the current module name (à la-top
) based on the filename passed, taking into account the proper-R
/-Q
options. For example, given-R Foo foolib
using-topfile foolib/bar.v
will set the module name toFoo.Bar
. CoqIDE now properly sets the module name for a given file based on its path (#8991, closes #8989, by Gaëtan Gilbert).Experimental: Coq flags and options can now be set on the command-line, e.g.
-set "Universe Polymorphism=true"
(#9876, by Gaëtan Gilbert).The
-native-compiler
flag ofcoqc
andcoqtop
now takes an argument which can have three values:no
disables native_computeyes
enables native_compute and precompiles.v
files to native codeondemand
enables native_compute but compiles code only whennative_compute
is called
The default value is
ondemand
. Note that this flag now has priority over the configure flag of the same name.A new
-bytecode-compiler
flag forcoqc
andcoqtop
controls whether conversion can use the VM. The default value isyes
.(#8870, by Maxime Dénès)
The pretty timing diff scripts (flag
TIMING=1
to acoq_makefile
-madeMakefile
, alsotools/make-both-single-timing-files.py
,tools/make-both-time-files.py
, andtools/make-one-time-file.py
) now correctly support non-UTF-8 characters in the output ofcoqc
/make
as well as printing to stdout, on both python2 and python3 (#9872, closes #9767 and #9705, by Jason Gross)coq_makefile's install target now errors if any file to install is missing (#9906, by Gaëtan Gilbert).
Preferences from
coqide.keys
are no longer overridden by modifiers preferences incoqiderc
(#10014, by Hugo Herbelin).
Specification language, type inference:
Fixing a missing check in interpreting instances of existential variables that are bound to local definitions. Might exceptionally induce an overhead if the cost of checking the conversion of the corresponding definitions is additionally high (#8217, closes #8215, by Hugo Herbelin).
A few improvements in inference of the return clause of
match
that can exceptionally introduce incompatibilities. This can be solved by writing an explicitreturn
clause, sometimes even simply an explicitreturn _
clause (#262, by Hugo Herbelin).Using non-projection values with the projection syntax is not allowed. For instance
0.(S)
is not a valid way to writeS 0
. Projections from non-primitive (emulated) records are allowed with warning "nonprimitive-projection-syntax" (#8829, by Gaëtan Gilbert).An option and attributes to control the automatic decision to declare an inductive type as template polymorphic were added. Warning "auto-template" (off by default) can trigger when an inductive is automatically declared template polymorphic without the attribute.
Inductive types declared by Funind will never be template polymorphic.
(#8488, by Gaëtan Gilbert)
Notations:
New command
Declare Scope
to explicitly declare a scope name before any use of it. Implicit declaration of a scope at the time ofBind Scope
,Delimit Scope
,Undelimit Scope
, orNotation
is deprecated (#7135, by Hugo Herbelin).Various bugs have been fixed (e.g. #9214 on removing spurious parentheses on abbreviations shortening a strict prefix of an application, by Hugo Herbelin).
Number Notation
now support inductive types in the input to printing functions (e.g., numeral notations can be defined for terms containing things like@cons nat O O
), and parsing functions now fully normalize terms including parameters of constructors (so that, e.g., a numeral notation whose parsing function outputs a proof ofNat.gcd x y = 1
will no longer fail to parse due to containing the constantNat.gcd
in the parameter-argument ofeq_refl
) (#9874, closes #9840 and #9844, by Jason Gross).Deprecated compatibility notations have actually been removed. Uses of these notations are generally easy to fix thanks to the hint contained in the deprecation warning emitted by Coq 8.8 and 8.9. For projects that require more than a handful of such fixes, there is a script that will do it automatically, using the output of
coqc
(#8638, by Jason Gross).Allow inspecting custom grammar entries by
Print Custom Grammar
(#10061, fixes #9681, by Jasper Hugunin, review by Pierre-Marie Pédrot and Hugo Herbelin).
The quote plugin was removed. If some users are interested in maintaining this plugin externally, the Coq development team can provide assistance for extracting the plugin and setting up a new repository (#7894, by Maxime Dénès).
Ltac:
Tactic names are no longer allowed to clash, even if they are not defined in the same section. For example, the following is no longer accepted:
Ltac foo := idtac. Section S. Ltac foo := fail. End S.
(#8555, by Maxime Dénès).Names of existential variables occurring in Ltac functions (e.g.
?[n]
or?n
in terms - not in patterns) are now interpreted the same way as other variable names occurring in Ltac functions (#7309, by Hugo Herbelin).
Tactics:
Removed the deprecated
romega
tactic (#8419, by Maxime Dénès and Vincent Laporte).Hint declaration and removal should now specify a database (e.g.
Hint Resolve foo : database
). When the database name is omitted, the hint is added to thecore
database (as previously), but a deprecation warning is emitted (#8987, by Maxime Dénès).There are now tactics in
PreOmega.v
calledZ.div_mod_to_equations
,Z.quot_rem_to_equations
, andZ.to_euclidean_division_equations
(which combines thediv_mod
andquot_rem
variants) which allowlia
,nia
, etc to supportZ.div
andZ.modulo
(Z.quot
andZ.rem
, respectively), by posing the specifying equation forZ.div
andZ.modulo
before replacing them with atoms (#8062, by Jason Gross).The syntax of the
autoapply
tactic was fixed to conform with preexisting documentation: it now takes awith
clause instead of ausing
clause (#9524, closes #7632, by Théo Zimmermann).Modes are now taken into account by
typeclasses eauto
for local hypotheses (#9996, fixes #5752, by Maxime Dénès, review by Pierre-Marie Pédrot).New variant
change_no_check
ofchange
, usable as a documented replacement ofconvert_concl_no_check
(#10012, #10017, #10053, and #10059, by Hugo Herbelin and Paolo G. Giarrusso).The simplified value returned by
field_simplify
is not always a fraction anymore. When the denominator is1
, it returnsx
while previously it was returningx/1
. This change could break codes that were post-processing application offield_simplify
to get rid of thesex/1
(#9854, by Laurent Théry, with help from Michael Soegtrop, Maxime Dénès, and Vincent Laporte).
SSReflect:
Clear discipline made consistent across the entire proof language. Whenever a clear switch
{x..}
comes immediately before an existing proof context entry (used as a view, as a rewrite rule or as name for a new context entry) then such entry is cleared too.E.g. The following sentences are elaborated as follows (when H is an existing proof context entry):
=> {x..} H
->=> {x..H} H
=> {x..} /H
->=> /v {x..H}
rewrite {x..} H
->rewrite E {x..H}
(#9341, by Enrico Tassi).
inE
now expandsy in r x
whenr
is asimpl_rel
. New{pred T}
notation for apred T
alias in thepred_sort
coercion class, simplifiedpredType
interface:pred_class
andmkPredType
deprecated,{pred T}
andPredType
should be used instead.if c return t then ...
now expectsc
to be a variable bound int
. NewnonPropType
interface matching types that do _not_ have sortProp
. Newrelpre R f
definition for the preimage of a relation R under f (#9995, by Georges Gonthier).
Commands:
Binders for an
Instance
now act more like binders for aTheorem
. Names may not be repeated, and may not overlap with section variable names (#8820, closes #8791, by Jasper Hugunin).Removed the deprecated
Implicit Tactic
family of commands (#8779, by Pierre-Marie Pédrot).The
Automatic Introduction
option has been removed and is now the default (#9001, by Emilio Jesús Gallego Arias).Arguments
now accepts names for arguments provided withextra_scopes
(#9117, by Maxime Dénès).The naming scheme for anonymous binders in a
Theorem
has changed to avoid conflicts with explicitly named binders (#9160, closes #8819, by Jasper Hugunin).Computation of implicit arguments now properly handles local definitions in the binders for an
Instance
, and can be mixed with implicit binders{x : T}
(#9307, closes #9300, by Jasper Hugunin).Declare Instance
now requires an instance name.The flag
Refine Instance Mode
has been turned off by default, meaning thatInstance
no longer opens a proof when a body is provided. The flag has been deprecated and will be removed in the next version.Command
Instance
, when no body is provided, now always opens a proof. This is a breaking change, as instance ofInstance ident1 : ident2.
whereident2
is a trivial class will have to be changed intoInstance ident1 : ident2 := {}.
orInstance ident1 : ident2. Proof. Qed.
(#9274, by Maxime Dénès).The flag
Program Mode
now means that theProgram
attribute is enabled for all commands that support it. In particular, it does not have any effect on tactics anymore. May cause some incompatibilities (#9410, by Maxime Dénès).The algorithm computing implicit arguments now behaves uniformly for primitive projection and application nodes (#9509, closes #9508, by Pierre-Marie Pédrot).
Hypotheses
andVariables
can now take implicit binders inside sections (#9364, closes #9363, by Jasper Hugunin).Removed deprecated option
Automatic Coercions Import
(#8094, by Maxime Dénès).The
Show Script
command has been deprecated (#9829, by Vincent Laporte).Coercion
does not warn ambiguous paths which are obviously convertible with existing ones. The ambiguous paths messages have been turned to warnings, thus now they could appear in the output ofcoqc
. The convertibility checking procedure for coercion paths is complete for paths consisting of coercions satisfying the uniform inheritance condition, but some coercion paths could be reported as ambiguous even if they are convertible with existing ones when they have coercions that don't satisfy the uniform inheritance condition (#9743, closes #3219, by Kazuhiko Sakaguchi).A new flag
Fast Name Printing
has been introduced. It changes the algorithm used for allocating bound variable names for a faster but less clever one (#9078, by Pierre-Marie Pédrot).Option
Typeclasses Axioms Are Instances
(compatibility option introduced in the previous version) is deprecated. UseDeclare Instance
for axioms which should be instances (#8920, by Gaëtan Gilbert).Removed option
Printing Primitive Projection Compatibility
(#9306, by Gaëtan Gilbert).
Standard Library:
Added
Bvector.BVeq
that decides whether twoBvector
s are equal. Added notations forBVxor
,BVand
,BVor
,BVeq
andBVneg
(#8171, by Yishuai Li).Added
ByteVector
type that can convert to and fromstring
(#8365, by Yishuai Li).Added lemmas about monotonicity of
N.double
andN.succ_double
, and about the upper bound of number represented by a vector. Allowed implicit vector length argument inNdigits.Bv2N
(#8815, by Yishuai Li).The prelude used to be automatically Exported and is now only Imported. This should be relevant only when importing files which don't use
-noinit
into files which do (#9013, by Gaëtan Gilbert).Added
Coq.Structures.OrderedTypeEx.String_as_OT
to make strings an ordered type, using lexical order (#7221, by Li Yao).Added lemmas about
Z.testbit
,Z.ones
, andZ.modulo
(#9425, by Andres Erbsen).Moved the
auto
hints of theFSet
library into a newfset
database (#9725, by Frédéric Besson).Added
Coq.Structures.EqualitiesFacts.PairUsualDecidableTypeFull
(#9984, by Jean-Christophe Léchenet and Oliver Nash).
Some error messages that show problems with a pair of non-matching values will now highlight the differences (#8669, by Jim Fehrle).
Changelog has been moved from a specific file
CHANGES.md
to the reference manual; former Credits chapter of the reference manual has been split in two parts: a History chapter which was enriched with additional historical information about Coq versions 1 to 5, and a Changes chapter which was enriched with the content formerly inCHANGES.md
andCOMPATIBILITY
(#9133, #9668, #9939, #9964, and #10085, by Théo Zimmermann, with help and ideas from Emilio Jesús Gallego Arias, Gaëtan Gilbert, Clément Pit-Claudel, Matthieu Sozeau, and Enrico Tassi).
Changes in 8.10+beta2¶
Many bug fixes and documentation improvements, in particular:
Tactics
Make the
discriminate
tactic work together withUniverse Polymorphism
and equality inType
. This, in particular, makesdiscriminate
compatible with the HoTT library https://github.com/HoTT/HoTT (#10205, by Andreas Lynge, review by Pierre-Marie Pédrot and Matthieu Sozeau).
SSReflect
Make the
case E: t
tactic work together withUniverse Polymorphism
and equality inType
. This makescase
compatible with the HoTT library https://github.com/HoTT/HoTT (#10302, fixes #10301, by Andreas Lynge, review by Enrico Tassi)Make the
rewrite /t
tactic work together withUniverse Polymorphism
. This makesrewrite
compatible with the HoTT library https://github.com/HoTT/HoTT (#10305, fixes #9336, by Andreas Lynge, review by Enrico Tassi)
CoqIDE
Fix CoqIDE instability on Windows after the update to gtk3 (#10360, by Michael Soegtrop, closes #9885).
Miscellaneous
Changes in 8.10+beta3¶
Kernel
Fix soundness issue with template polymorphism (#9294).
Declarations of template-polymorphic inductive types ignored the provenance of the universes they were abstracting on and did not detect if they should be greater or equal to \(\Set\) in general. Previous universes and universes introduced by the inductive definition could have constraints that prevented their instantiation with e.g. \(\Prop\), resulting in unsound instantiations later. The implemented fix only allows abstraction over universes introduced by the inductive declaration, and properly records all their constraints by making them by default only \(>= \Prop\). It is also checked that a template polymorphic inductive actually is polymorphic on at least one universe.
This prevents inductive declarations in sections to be universe polymorphic over section parameters. For a backward compatible fix, simply hoist the inductive definition out of the section. An alternative is to declare the inductive as universe-polymorphic and cumulative in a universe-polymorphic section: all universes and constraints will be properly gathered in this case. See Template polymorphism for a detailed exposition of the rules governing template-polymorphic types.
To help users incrementally fix this issue, a command line option
-no-template-check
and a global flagTemplate Check
are available to selectively disable the new check. Use at your own risk.(#9918, by Matthieu Sozeau and Maxime Dénès).
User messages
Improve the ambiguous paths warning to indicate which path is ambiguous with new one (#10336, closes #3219, by Kazuhiko Sakaguchi).
Extraction
Fix extraction to OCaml of primitive machine integers; see Primitive Integers (#10430, fixes #10361, by Vincent Laporte).
Fix a printing bug of OCaml extraction on dependent record projections, which produced improper
assert false
. This change makes the OCaml extractor internally inline record projections by default; thus the monolithic OCaml extraction (Extraction
andRecursive Extraction
) does not produce record projection constants anymore except for record projections explicitly instructed to extract, and records declared in opaque modules (#10577, fixes #7348, by Kazuhiko Sakaguchi).
Standard library
Changes in 8.10.0¶
Changes in 8.10.1¶
A few bug fixes and documentation improvements, in particular:
Kernel
Fix proof of False when using \(\SProp\) (incorrect De Bruijn handling when inferring the relevance mark of a function) (#10904, by Pierre-Marie Pédrot).
Tactics
Tactic language
CoqIDE
Extraction
Changes in 8.10.2¶
Kernel
Fixed a critical bug of template polymorphism and nonlinear universes (#11128, fixes #11039, by Gaëtan Gilbert).
Fixed an anomaly “Uncaught exception Constr.DestKO” on
Inductive
(#11052, fixes #11048, by Gaëtan Gilbert).Fixed an anomaly “not enough abstractions in fix body” (#11014, fixes #8459, by Gaëtan Gilbert).
Notations
Fixed an 8.10 regression related to the printing of coercions associated with notations (#11090, fixes #11033, by Hugo Herbelin).
CoqIDE
Fixed uneven dimensions of CoqIDE panels when window has been resized (#11070, fixes 8.10-regression #10956, by Guillaume Melquiond).
Do not include final stops in queries (#11069, fixes 8.10-regression #11058, by Guillaume Melquiond).
Infrastructure and dependencies
Version 8.9¶
Summary of changes¶
Coq version 8.9 contains the result of refinements and stabilization of features and deprecations or removals of deprecated features, cleanups of the internals of the system and API along with a few new features. This release includes many user-visible changes, including deprecations that are documented in the next subsection and new features that are documented in the reference manual. Here are the most important changes:
Kernel: mutually recursive records are now supported, by Pierre-Marie Pédrot.
Notations:
Support for autonomous grammars of terms called “custom entries”, by Hugo Herbelin (see Section Custom entries of the reference manual).
Deprecated notations of the standard library will be removed in the next version of Coq, see the next subsection for a script to ease porting, by Jason Gross and Jean-Christophe Léchenet.
Added the
Number Notation
command for registering decimal numeral notations for custom types, by Daniel de Rauglaudre, Pierre Letouzey and Jason Gross.
Tactics: Introduction tactics
intro
/intros
on a goal that is an existential variable now force a refinement of the goal into a dependent product rather than failing, by Hugo Herbelin.Decision procedures: deprecation of tactic
romega
in favor oflia
and removal offourier
, replaced bylra
which subsumes it, by Frédéric Besson, Maxime Dénès, Vincent Laporte and Laurent Théry.Proof language: focusing bracket
{
now supports named goals, e.g.[x]:{
will focus on a goal (existential variable) namedx
, by Théo Zimmermann.SSReflect: the implementation of delayed clear was simplified by Enrico Tassi: the variables are always renamed using inaccessible names when the clear switch is processed and finally cleared at the end of the intro pattern. In addition to that, the use-and-discard flag
{}
typical of rewrite rules can now be also applied to views, e.g.=> {}/v
appliesv
and then clearsv
. See Section Introduction in the context.Vernacular:
Experimental support for attributes on commands, by Vincent Laporte, as in
#[local] Lemma foo : bar.
Tactics and tactic notations now support thedeprecated
attribute.Removed deprecated commands
Arguments Scope
andImplicit Arguments
in favor ofArguments
, with the help of Jasper Hugunin.New flag
Uniform Inductive Parameters
by Jasper Hugunin to avoid repeating uniform parameters in constructor declarations.New commands
Hint Variables
andHint Constants
, by Matthieu Sozeau, for controlling the opacity status of variables and constants in hint databases. It is recommended to always use these commands after creating a hint database withCreate HintDb
.Multiple sections with the same name are now allowed, by Jasper Hugunin.
Library: additions and changes in the
VectorDef
,Ascii
, andString
libraries. Syntax notations are now available only when usingImport
of libraries and not merelyRequire
, by various contributors (source of incompatibility, see the next subsection for details).Toplevels:
coqtop
andcoqide
can now display diffs between proof steps in color, using theDiffs
option, by Jim Fehrle.Documentation: we integrated a large number of fixes to the new Sphinx documentation by various contributors, coordinated by Clément Pit-Claudel and Théo Zimmermann.
Tools: removed the
gallina
utility and the homebrewedEmacs
mode.Packaging: as in Coq 8.8.2, the Windows installer now includes many more external packages that can be individually selected for installation, by Michael Soegtrop.
Version 8.9 also comes with a bunch of smaller-scale changes and improvements regarding the different components of the system. Most important ones are documented in the next subsection file.
On the implementation side, the dev/doc/changes.md
file documents
the numerous changes to the implementation and improvements of
interfaces. The file provides guidelines on porting a plugin to the new
version and a plugin development tutorial kept in sync with Coq was
introduced by Yves Bertot http://github.com/ybertot/plugin_tutorials.
The new dev/doc/critical-bugs
file documents the known critical bugs
of Coq and affected releases.
The efficiency of the whole system has seen improvements thanks to contributions from Gaëtan Gilbert, Pierre-Marie Pédrot, and Maxime Dénès.
Maxime Dénès, Emilio Jesús Gallego Arias, Gaëtan Gilbert, Michael Soegtrop, Théo Zimmermann worked on maintaining and improving the continuous integration system.
The OPAM repository for Coq packages has been maintained by Guillaume Melquiond, Matthieu Sozeau, Enrico Tassi with contributions from many users. A list of packages is available at https://coq.inria.fr/opam/www/.
The 54 contributors for this version are Léo Andrès, Rin Arakaki, Benjamin Barenblat, Langston Barrett, Siddharth Bhat, Martin Bodin, Simon Boulier, Timothy Bourke, Joachim Breitner, Tej Chajed, Arthur Charguéraud, Pierre Courtieu, Maxime Dénès, Andres Erbsen, Jim Fehrle, Julien Forest, Emilio Jesus Gallego Arias, Gaëtan Gilbert, Matěj Grabovský, Jason Gross, Samuel Gruetter, Armaël Guéneau, Hugo Herbelin, Jasper Hugunin, Ralf Jung, Sam Pablo Kuper, Ambroise Lafont, Leonidas Lampropoulos, Vincent Laporte, Peter LeFanu Lumsdaine, Pierre Letouzey, Jean-Christophe Léchenet, Nick Lewycky, Yishuai Li, Sven M. Hallberg, Assia Mahboubi, Cyprien Mangin, Guillaume Melquiond, Perry E. Metzger, Clément Pit-Claudel, Pierre-Marie Pédrot, Daniel R. Grayson, Kazuhiko Sakaguchi, Michael Soegtrop, Matthieu Sozeau, Paul Steckler, Enrico Tassi, Laurent Théry, Anton Trunov, whitequark, Théo Winterhalter, Zeimer, Beta Ziliani, Théo Zimmermann.
Many power users helped to improve the design of the new features via the issue and pull request system, the Coq development mailing list or the coq-club@inria.fr mailing list. It would be impossible to mention exhaustively the names of everybody who to some extent influenced the development.
Version 8.9 is the fourth release of Coq developed on a time-based development cycle. Its development spanned 7 months from the release of Coq 8.8. The development moved to a decentralized merging process during this cycle. Guillaume Melquiond was in charge of the release process and is the maintainer of this release. This release is the result of ~2,000 commits and ~500 PRs merged, closing 75+ issues.
The Coq development team welcomed Vincent Laporte, a new Coq engineer working with Maxime Dénès in the Coq consortium.
Details of changes in 8.9+beta1¶
Kernel
Mutually defined records are now supported.
Notations
New support for autonomous grammars of terms, called "custom entries" (see chapter "Syntax extensions" of the reference manual).
Deprecated compatibility notations will actually be removed in the next version of Coq. Uses of these notations are generally easy to fix thanks to the hint contained in the deprecation warnings. For projects that require more than a handful of such fixes, there is a script that will do it automatically, using the output of
coqc
. The script contains documentation on its usage in a comment at the top.
Tactics
Added toplevel goal selector
!
which expects a single focused goal. Use withSet Default Goal Selector
to force focusing before tactics are called.The undocumented "nameless" forms
fix N
,cofix
that were deprecated in 8.8 have been removed from Ltac's syntax; please usefix ident N/cofix ident
to explicitly name the (co)fixpoint hypothesis to be introduced.Introduction tactics
intro
/intros
on a goal that is an existential variable now force a refinement of the goal into a dependent product rather than failing.Support for
fix
/cofix
added in Ltacmatch
andlazymatch
.Ltac backtraces now include trace information about tactics called by OCaml-defined tactics.
Option
Ltac Debug
now applies also to terms built using Ltac functions.Deprecated the
Implicit Tactic
family of commands.The default program obligation tactic uses a bounded proof search instead of an unbounded and potentially non-terminating one now (source of incompatibility).
The
simple apply
tactic now respects theOpaque
flag when called from Ltac (auto
still does not respect it).Tactic
constr_eq
now adds universe constraints needed for the identity to the context (it used to ignore them). New tacticconstr_eq_strict
checks that the required constraints already hold without adding new ones. Preexisting tacticconstr_eq_nounivs
can still be used if you really want to ignore universe constraints.Tactics and tactic notations now understand the
deprecated
attribute.The
fourier
tactic has been removed. Please now uselra
instead. You may need to addRequire Import Lra
to your developments. For compatibility, we now definefourier
as a deprecated alias oflra
.The
romega
tactics have been deprecated; please uselia
instead.
Focusing
Focusing bracket
{
now supports named goal selectors, e.g.[x]: {
will focus on a goal (existential variable) namedx
. As usual, unfocus with}
once the subgoal is fully solved.
Specification language
A fix to unification (which was sensitive to the ascii name of variables) may occasionally change type inference in incompatible ways, especially regarding the inference of the return clause of
match
.
Standard Library
Added
Ascii.eqb
andString.eqb
and the=?
notation for them, and proved some lemmas about them. Note that this might cause incompatibilities if you have, e.g.,string_scope
andZ_scope
both open withstring_scope
on top, and expect=?
to refer toZ.eqb
. Solution: wrap_ =? _
in(_ =? _)%Z
(or whichever scope you want).Added
Ndigits.N2Bv_sized
, and proved some lemmas about it. DeprecatedNdigits.N2Bv_gen
.The scopes
int_scope
anduint_scope
have been renamed todec_int_scope
anddec_uint_scope
, to clash less with ssreflect and other packages. They are still delimited by%int
and%uint
.Syntax notations for
string
,ascii
,Z
,positive
,N
,R
, andint31
are no longer available merely byRequire
ing the files that define the inductives. You mustImport
Coq.Strings.String.StringSyntax
(afterRequire
Coq.Strings.String
),Coq.Strings.Ascii.AsciiSyntax
(afterRequire
Coq.Strings.Ascii
),Coq.ZArith.BinIntDef
,Coq.PArith.BinPosDef
,Coq.NArith.BinNatDef
,Coq.Reals.Rdefinitions
, andCoq.Numbers.Cyclic.Int31.Int31
, respectively, to be able to use these notations. Note that passing-compat 8.8
or issuingRequire Import Coq.Compat.Coq88
will make these notations available. Users wishing to port their developments automatically may downloadfix.py
from https://gist.github.com/JasonGross/5d4558edf8f5c2c548a3d96c17820169 and run a command likewhile true; do make -Okj 2>&1 | /path/to/fix.py; done
and get a cup of coffee. (This command must be manually interrupted once the build finishes all the way though. Note also that this method is not fail-proof; you may have to adjust some scopes if you were relying on string notations not being available even whenstring_scope
was open.)Numeral syntax for
nat
is no longer available without loading the entire prelude (Require Import Coq.Init.Prelude
). This only impacts users running Coq without the init library (-nois
or-noinit
) and also issuingRequire Import Coq.Init.Datatypes
.
Tools
Coq_makefile lets one override or extend the following variables from the command line:
COQFLAGS
,COQCHKFLAGS
,COQDOCFLAGS
.COQFLAGS
is now entirely separate fromCOQLIBS
, so in custom Makefiles$(COQFLAGS)
should be replaced by$(COQFLAGS) $(COQLIBS)
.Removed the
gallina
utility (extracts specification from Coq vernacular files). If you would like to maintain this tool externally, please contact us.Removed the Emacs modes distributed with Coq. You are advised to use Proof-General (and optionally Company-Coq) instead. If your use case is not covered by these alternative Emacs modes, please open an issue. We can help set up external maintenance as part of Proof-General, or independently as part of coq-community.
Commands
Removed deprecated commands
Arguments Scope
andImplicit Arguments
(not the option). Use theArguments
command instead.Nested proofs may be enabled through the option
Nested Proofs Allowed
. By default, they are disabled and produce an error. The deprecation warning which used to occur when using nested proofs has been removed.Added option
Uniform Inductive Parameters
which abstracts over parameters before typechecking constructors, allowing to write for exampleInductive list (A : Type) := nil : list | cons : A -> list -> list.
New
Set Hint Variables/Constants Opaque/Transparent
commands for setting globally the opacity flag of variables and constants in hint databases, overriding the opacity setting of the hint database.Added generic syntax for "attributes", as in:
#[local] Lemma foo : bar.
Added the
Numeral Notation
command for registering decimal numeral notations for custom typesThe
Set SsrHave NoTCResolution
command no longer has special global scope. If you want the previous behavior, useGlobal Set SsrHave NoTCResolution
.Multiple sections with the same name are allowed.
Coq binaries and process model
Before 8.9, Coq distributed a single
coqtop
binary and a set of dynamically loadable plugins that used to take over the main loop for tasks such as IDE language server or parallel proof checking.These plugins have been turned into full-fledged binaries so each different process has associated a particular binary now, in particular
coqidetop
is the CoqIDE language server, andcoq{proof,tactic,query}worker
are in charge of task-specific and parallel proof checking.
SSReflect
The implementation of delayed clear switches in intro patterns is now simpler to explain:
The immediate effect of a clear switch like
{x}
is to rename the variablex
to_x_
(i.e. a reserved identifier that cannot be mentioned explicitly)The delayed effect of
{x}
is that_x_
is cleared at the end of the intro patternA clear switch immediately before a view application like
{x}/v
is translated to/v{x}
.
In particular, the third rule lets one write
{x}/v
even ifv
uses the variablex
: indeed the view is executed before the renaming.An empty clear switch is now accepted in intro patterns before a view application whenever the view is a variable. One can now write
{}/v
to mean{v}/v
. Remark that{}/x
is very similar to the idiom{}e
for the rewrite tactic (the equatione
is used for rewriting and then discarded).
Standard Library
There are now conversions between
string
andpositive
,Z
,nat
, andN
in binary, octal, and hex.
Display diffs between proof steps
coqtop
andcoqide
can now highlight the differences between proof steps in color. This can be enabled from the command line or theSet Diffs "on"/"off"/"removed"
command. Please see the documentation for details. Showing diffs in Proof General requires small changes to PG (under discussion).
Notations
Added
++
infix forVectorDef.append
. Note that this might cause incompatibilities if you have, e.g.,list_scope
andvector_scope
both open withvector_scope
on top, and expect++
to refer toapp
. Solution: wrap_ ++ _
in(_ ++ _)%list
(or whichever scope you want).
Changes in 8.8.0¶
Various bug fixes.
Changes in 8.8.1¶
Some quality-of-life fixes.
Numerous improvements to the documentation.
Fix a critical bug related to primitive projections and
native_compute
.Ship several additional Coq libraries with the Windows installer.
Version 8.8¶
Summary of changes¶
Coq version 8.8 contains the result of refinements and stabilization of features and deprecations, cleanups of the internals of the system along with a few new features. The main user visible changes are:
Kernel: fix a subject reduction failure due to allowing fixpoints on non-recursive values, by Matthieu Sozeau. Handling of evars in the VM (the kernel still does not accept evars) by Pierre-Marie Pédrot.
Notations: many improvements on recursive notations and support for destructuring patterns in the syntax of notations by Hugo Herbelin.
Proof language: tacticals for profiling, timing and checking success or failure of tactics by Jason Gross. The focusing bracket
{
supports single-numbered goal selectors, e.g.2:{
, by Théo Zimmermann.Vernacular: deprecation of commands and more uniform handling of the
Local
flag, by Vincent Laporte and Maxime Dénès, part of a larger attribute system overhaul. ExperimentalShow Extraction
command by Pierre Letouzey. Coercion now acceptsProp
orType
as a source by Arthur Charguéraud.Export
modifier for options allowing to export the option to modules thatImport
and not onlyRequire
a module, by Pierre-Marie Pédrot.Universes: many user-level and API level enhancements: qualified naming and printing, variance annotations for cumulative inductive types, more general constraints and enhancements of the minimization heuristics, interaction with modules by Gaëtan Gilbert, Pierre-Marie Pédrot and Matthieu Sozeau.
Library: Decimal Numbers library by Pierre Letouzey and various small improvements.
Documentation: a large community effort resulted in the migration of the reference manual to the Sphinx documentation tool. The result is this manual. The new documentation infrastructure (based on Sphinx) is by Clément Pit-Claudel. The migration was coordinated by Maxime Dénès and Paul Steckler, with some help of Théo Zimmermann during the final integration phase. The 14 people who ported the manual are Calvin Beck, Heiko Becker, Yves Bertot, Maxime Dénès, Richard Ford, Pierre Letouzey, Assia Mahboubi, Clément Pit-Claudel, Laurence Rideau, Matthieu Sozeau, Paul Steckler, Enrico Tassi, Laurent Théry, Nikita Zyuzin.
Tools: experimental
-mangle-names
option tocoqtop
/coqc
for linting proof scripts, by Jasper Hugunin.
On the implementation side, the dev/doc/changes.md
file
documents the numerous changes to the implementation and improvements of
interfaces. The file provides guidelines on porting a plugin to the new
version.
Version 8.8 also comes with a bunch of smaller-scale changes and improvements regarding the different components of the system. Most important ones are documented in the next subsection file.
The efficiency of the whole system has seen improvements thanks to contributions from Gaëtan Gilbert, Pierre-Marie Pédrot, Maxime Dénès and Matthieu Sozeau and performance issue tracking by Jason Gross and Paul Steckler.
The official wiki and the bugtracker of Coq migrated to the GitHub platform, thanks to the work of Pierre Letouzey and Théo Zimmermann. Gaëtan Gilbert, Emilio Jesús Gallego Arias worked on maintaining and improving the continuous integration system.
The OPAM repository for Coq packages has been maintained by Guillaume Melquiond, Matthieu Sozeau, Enrico Tassi with contributions from many users. A list of packages is available at https://coq.inria.fr/opam/www/.
The 44 contributors for this version are Yves Bertot, Joachim Breitner, Tej Chajed, Arthur Charguéraud, Jacques-Pascal Deplaix, Maxime Dénès, Jim Fehrle, Julien Forest, Yannick Forster, Gaëtan Gilbert, Jason Gross, Samuel Gruetter, Thomas Hebb, Hugo Herbelin, Jasper Hugunin, Emilio Jesus Gallego Arias, Ralf Jung, Johannes Kloos, Matej Košík, Robbert Krebbers, Tony Beta Lambda, Vincent Laporte, Peter LeFanu Lumsdaine, Pierre Letouzey, Farzon Lotfi, Cyprien Mangin, Guillaume Melquiond, Raphaël Monat, Carl Patenaude Poulin, Pierre-Marie Pédrot, Clément Pit-Claudel, Matthew Ryan, Matt Quinn, Sigurd Schneider, Bernhard Schommer, Michael Soegtrop, Matthieu Sozeau, Arnaud Spiwack, Paul Steckler, Enrico Tassi, Anton Trunov, Martin Vassor, Vadim Zaliva and Théo Zimmermann.
Version 8.8 is the third release of Coq developed on a time-based development cycle. Its development spanned 6 months from the release of Coq 8.7 and was based on a public roadmap. The development process was coordinated by Matthieu Sozeau. Maxime Dénès was in charge of the release process. Théo Zimmermann is the maintainer of this release.
Many power users helped to improve the design of the new features via the bug tracker, the pull request system, the Coq development mailing list or the coq-club@inria.fr mailing list. Special thanks to the users who contributed patches and intensive brain-storming and code reviews, starting with Jason Gross, Ralf Jung, Robbert Krebbers and Amin Timany. It would however be impossible to mention exhaustively the names of everybody who to some extent influenced the development.
The Coq consortium, an organization directed towards users and supporters of the system, is now running and employs Maxime Dénès. The contacts of the Coq Consortium are Yves Bertot and Maxime Dénès.
Details of changes in 8.8+beta1¶
Kernel
Support for template polymorphism for definitions was removed. May trigger more "universe inconsistency" errors in rare occasions.
Fixpoints are no longer allowed on non-recursive inductive types.
Notations
Recursive notations with the recursive pattern repeating on the right (e.g. "( x ; .. ; y ; z )") now supported.
Notations with a specific level for the leftmost nonterminal, when printing-only, are supported.
Notations can now refer to the syntactic category of patterns (as in "fun 'pat =>" or "match p with pat => ... end"). Two variants are available, depending on whether a single variable is considered as a pattern or not.
Recursive notations now support ".." patterns with several occurrences of the recursive term or binder, possibly mixing terms and binders, possibly in reverse left-to-right order.
"Locate" now working also on notations of the form "x + y" (rather than "_ + _").
Specification language
When printing clauses of a "match", clauses with same right-hand side are factorized and the last most factorized clause with no variables, if it exists, is turned into a default clause. Use "Unset Printing Allow Default Clause" do deactivate printing of a default clause. Use "Unset Printing Factorizable Match Patterns" to deactivate factorization of clauses with same right-hand side.
Tactics
On Linux, "native_compute" calls can be profiled using the "perf" utility. The command "Set NativeCompute Profiling" enables profiling, and "Set NativeCompute Profile Filename" customizes the profile filename.
The tactic "omega" is now aware of the bodies of context variables such as "x := 5 : Z" (see #1362). This could be disabled via Unset Omega UseLocalDefs.
The tactic "romega" is also aware now of the bodies of context variables.
The tactic "zify" resp. "omega with N" is now aware of N.pred.
Tactic "decide equality" now able to manage constructors which contain proofs.
Added tactics reset ltac profile, show ltac profile (and variants)
Added tactics restart_timer, finish_timing, and time_constr as an experimental way of timing Ltac's evaluation phase
Added tactic optimize_heap, analogous to the Vernacular Optimize Heap, which performs a major garbage collection and heap compaction in the OCaml run-time system.
The tactics "dtauto", "dintuition", "firstorder" now handle inductive types with let bindings in the parameters.
The tactic
dtauto
now handles some inductives such as@sigT A (fun _ => B)
as non-dependent conjunctions.A bug fixed in
rewrite H in *
andrewrite H in * |-
may cause a few rare incompatibilities (it was unintendedly recursively rewriting in the side conditions generated by H).Added tactics "assert_succeeds tac" and "assert_fails tac" to ensure properties of the execution of a tactic without keeping the effect of the execution.
vm_compute
now supports existential variables.Calls to
shelve
andgive_up
within calls to tacticrefine
now working.Deprecated tactic
appcontext
was removed.
Focusing
Focusing bracket
{
now supports single-numbered goal selector, e.g.2: {
will focus on the second subgoal. As usual, unfocus with}
once the subgoal is fully solved. TheFocus
andUnfocus
commands are now deprecated.
Commands
Proofs ending in "Qed exporting ident, .., ident" are not supported anymore. Constants generated during
abstract
are kept private to the local environment.The deprecated Coercion Local, Open Local Scope, Notation Local syntax was removed. Use Local as a prefix instead.
For the Extraction Language command, "OCaml" is spelled correctly. The older "Ocaml" is still accepted, but deprecated.
Using “Require” inside a section is deprecated.
An experimental command "Show Extraction" allows to extract the content of the current ongoing proof (grant wish #4129).
Coercion now accepts the type of its argument to be "Prop" or "Type".
The "Export" modifier can now be used when setting and unsetting options, and will result in performing the same change when the module corresponding the command is imported.
The
Axiom
command does not automatically declare axioms as instances when their type is a class. Previous behavior can be restored usingSet Typeclasses Axioms Are Instances
.
Universes
Qualified naming of global universes now works like other namespaced objects (e.g. constants), with a separate namespace, inside and across module and library boundaries. Global universe names introduced in an inductive / constant / Let declaration get qualified with the name of the declaration.
Universe cumulativity for inductive types is now specified as a variance for each polymorphic universe. See the reference manual for more information.
Inference of universe constraints with cumulative inductive types produces more general constraints. Unsetting new option Cumulativity Weak Constraints produces even more general constraints (but may produce too many universes to be practical).
Fix #5726: Notations that start with
Type
now support universe instances with@{u}
.with Definition
now understands universe declarations (like@{u| Set < u}
).
Tools
Coq can now be run with the option -mangle-names to change the auto-generated name scheme. This is intended to function as a linter for developments that want to be robust to changes in auto-generated names. This feature is experimental, and may change or disappear without warning.
GeoProof support was removed.
Checker
The checker now accepts filenames in addition to logical paths.
CoqIDE
Find and Replace All report the number of occurrences found; Find indicates when it wraps.
coqdep
Learned to read -I, -Q, -R and filenames from _CoqProject files. This is used by coq_makefile when generating dependencies for .v files (but not other files).
Documentation
The Coq FAQ, formerly located at https://coq.inria.fr/faq, has been moved to the GitHub wiki section of this repository; the main entry page is https://github.com/coq/coq/wiki/The-Coq-FAQ.
Documentation: a large community effort resulted in the migration of the reference manual to the Sphinx documentation tool. The result is partially integrated in this version.
Standard Library
New libraries Coq.Init.Decimal, Coq.Numbers.DecimalFacts, Coq.Numbers.DecimalNat, Coq.Numbers.DecimalPos, Coq.Numbers.DecimalN, Coq.Numbers.DecimalZ, Coq.Numbers.DecimalString providing a type of decimal numbers, some facts about them, and conversions between decimal numbers and nat, positive, N, Z, and string.
Added [Coq.Strings.String.concat] to concatenate a list of strings inserting a separator between each item
Notation
'
for Zpos in QArith was removed.Some deprecated aliases are now emitting warnings when used.
Compatibility support
Support for compatibility with versions before 8.6 was dropped.
Options
The following deprecated options have been removed:
Refolding Reduction
Standard Proposition Elimination
Dependent Propositions Elimination
Discriminate Introduction
Shrink Abstract
Tactic Pattern Unification
Intuition Iff Unfolding
Injection L2R Pattern Order
Record Elimination Schemes
Match Strict
Tactic Compat Context
Typeclasses Legacy Resolution
Typeclasses Module Eta
Typeclass Resolution After Apply
Details of changes in 8.8.0¶
Tools
Asynchronous proof delegation policy was fixed. Since version 8.7 Coq was ignoring previous runs and the
-async-proofs-delegation-threshold
option did not have the expected behavior.
Tactic language
The undocumented "nameless" forms
fix N
,cofix
have been deprecated; please usefix ident N /cofix ident
to explicitly name the (co)fixpoint hypothesis to be introduced.
Documentation
The reference manual is now fully ported to Sphinx.
Other small deprecations and bug fixes.
Details of changes in 8.8.1¶
Kernel
Fix a critical bug with cofixpoints and
vm_compute
/native_compute
(#7333).Fix a critical bug with modules and algebraic universes (#7695)
Fix a critical bug with inlining of polymorphic constants (#7615).
Fix a critical bug with universe polymorphism and
vm_compute
(#7723). Was present since 8.5.
Notations
Fixed unexpected collision between only-parsing and only-printing notations (issue #7462).
Windows installer
The Windows installer now includes external packages Ltac2 and Equations (it included the Bignums package since 8.8+beta1).
Many other bug fixes, documentation improvements (including fixes of regressions due to the Sphinx migration), and user message improvements (for details, see the 8.8.1 milestone at https://github.com/coq/coq/milestone/13?closed=1).
Details of changes in 8.8.2¶
Documentation
A PDF version of the reference manual is available once again.
Tools
The coq-makefile targets
print-pretty-timed
,print-pretty-timed-diff
, andprint-pretty-single-time-diff
now correctly label the "before" and "after" columns, rather than swapping them.
Kernel
The kernel does not tolerate capture of global universes by polymorphic universe binders, fixing a soundness break (triggered only through custom plugins)
Windows installer
The Windows installer now includes many more external packages that can be individually selected for installation.
Many other bug fixes and lots of documentation improvements (for details, see the 8.8.2 milestone at https://github.com/coq/coq/milestone/15?closed=1).
Version 8.7¶
Summary of changes¶
Coq version 8.7 contains the result of refinements, stabilization of features and cleanups of the internals of the system along with a few new features. The main user visible changes are:
New tactics: variants of tactics supporting existential variables
eassert
,eenough
, etc... by Hugo Herbelin. Tacticsextensionality in H
andinversion_sigma
by Jason Gross,specialize with ...
accepting partial bindings by Pierre Courtieu.Cumulative Polymorphic Inductive
types, allowing cumulativity of universes to go through applied inductive types, by Amin Timany and Matthieu Sozeau.Integration of the SSReflect plugin and its documentation in the reference manual, by Enrico Tassi, Assia Mahboubi and Maxime Dénès.
The
coq_makefile
tool was completely redesigned to improve its maintainability and the extensibility of generated Makefiles, and to make_CoqProject
files more palatable to IDEs by Enrico Tassi.
Coq 8.7 involved a large amount of work on cleaning and speeding up the code
base, notably the work of Pierre-Marie Pédrot on making the tactic-level system
insensitive to existential variable expansion, providing a safer API to plugin
writers and making the code more robust. The dev/doc/changes.txt
file
documents the numerous changes to the implementation and improvements of
interfaces. An effort to provide an official, streamlined API to plugin writers
is in progress, thanks to the work of Matej Košík.
Version 8.7 also comes with a bunch of smaller-scale changes and improvements regarding the different components of the system. We shall only list a few of them.
The efficiency of the whole system has been significantly improved thanks to contributions from Pierre-Marie Pédrot, Maxime Dénès and Matthieu Sozeau and performance issue tracking by Jason Gross and Paul Steckler.
Thomas Sibut-Pinote and Hugo Herbelin added support for side effect hooks in cbv, cbn and simpl. The side effects are provided via a plugin available at https://github.com/herbelin/reduction-effects/.
The BigN, BigZ, BigQ libraries are no longer part of the Coq standard library, they are now provided by a separate repository https://github.com/coq/bignums, maintained by Pierre Letouzey.
In the Reals library, IZR
has been changed to produce a compact representation
of integers and real constants are now represented using IZR
(work by
Guillaume Melquiond).
Standard library additions and improvements by Jason Gross, Pierre Letouzey and others, documented in the next subsection file.
The mathematical proof language/declarative mode plugin was removed from the archive.
The OPAM repository for Coq packages has been maintained by Guillaume Melquiond, Matthieu Sozeau, Enrico Tassi with contributions from many users. A list of packages is available at https://coq.inria.fr/opam/www/.
Packaging tools and software development kits were prepared by Michael Soegtrop with the help of Maxime Dénès and Enrico Tassi for Windows, and Maxime Dénès for MacOS X. Packages are regularly built on the Travis continuous integration server.
The contributors for this version are Abhishek Anand, C.J. Bell, Yves Bertot, Frédéric Besson, Tej Chajed, Pierre Courtieu, Maxime Dénès, Julien Forest, Gaëtan Gilbert, Jason Gross, Hugo Herbelin, Emilio Jesús Gallego Arias, Ralf Jung, Matej Košík, Xavier Leroy, Pierre Letouzey, Assia Mahboubi, Cyprien Mangin, Erik Martin-Dorel, Olivier Marty, Guillaume Melquiond, Sam Pablo Kuper, Benjamin Pierce, Pierre-Marie Pédrot, Lars Rasmusson, Lionel Rieg, Valentin Robert, Yann Régis-Gianas, Thomas Sibut-Pinote, Michael Soegtrop, Matthieu Sozeau, Arnaud Spiwack, Paul Steckler, George Stelle, Pierre-Yves Strub, Enrico Tassi, Hendrik Tews, Amin Timany, Laurent Théry, Vadim Zaliva and Théo Zimmermann.
The development process was coordinated by Matthieu Sozeau with the help of Maxime Dénès, who was also in charge of the release process. Théo Zimmermann is the maintainer of this release.
Many power users helped to improve the design of the new features via the bug tracker, the pull request system, the Coq development mailing list or the Coq-Club mailing list. Special thanks to the users who contributed patches and intensive brain-storming and code reviews, starting with Jason Gross, Ralf Jung, Robbert Krebbers, Xavier Leroy, Clément Pit–Claudel and Gabriel Scherer. It would however be impossible to mention exhaustively the names of everybody who to some extent influenced the development.
Version 8.7 is the second release of Coq developed on a time-based development cycle. Its development spanned 9 months from the release of Coq 8.6 and was based on a public road-map. It attracted many external contributions. Code reviews and continuous integration testing were systematically used before integration of new features, with an important focus given to compatibility and performance issues, resulting in a hopefully more robust release than Coq 8.6 while maintaining compatibility.
Coq Enhancement Proposals (CEPs for short) and open pull request discussions were used to discuss publicly the new features.
The Coq consortium, an organization directed towards users and supporters of the system, is now upcoming and will rely on Inria’s newly created Foundation.
Potential compatibility issues¶
Extra superfluous names in introduction patterns may now raise an error rather than a warning when the superfluous name is already in use. The easy fix is to remove the superfluous name.
Details of changes in 8.7+beta1¶
Tactics
New tactic "extensionality in H" which applies (possibly dependent) functional extensionality in H supposed to be a quantified equality until giving a bare equality.
New tactic
inversion_sigma
which turns equalities of dependent pairs (e.g.,existT P x p = existT P y q
, frequently left over byinversion
on a dependent type family) into pairs of equalities (e.g., a hypothesisH : x = y
and a hypothesis of typerew H in p = q
); these hypotheses can subsequently be simplified usingsubst
, without ever invoking any kind of axiom asserting uniqueness of identity proofs. If you want to explicitly specify the hypothesis to be inverted, or name the generated hypotheses, you can invokeinduction H as [H1 H2] using eq_sigT_rect
. The tactic also works forsig
,sigT2
, andsig2
, and there are similareq_sig*_rect
induction lemmas.Tactic "specialize with ..." now accepts any partial bindings. Missing bindings are either solved by unification or left quantified in the hypothesis.
New representation of terms that statically ensure stability by evar-expansion. This has several consequences.
In terms of performance, this adds a cost to every term destructuration, but at the same time most eager evar normalizations were removed, which couterbalances this drawback and even sometimes outperforms the old implementation. For instance, many operations that would require O(n) normalization of the term are now O(1) in tactics. YMMV.
This triggers small changes in unification, which was not evar-insensitive. Most notably, the new implementation recognizes Miller patterns that were missed before because of a missing normalization step. Hopefully this should be fairly uncommon.
Tactic "auto with real" can now discharge comparisons of literals.
The types of variables in patterns of "match" are now beta-iota-reduced after type checking. This has an impact on the type of the variables that the tactic "refine" introduces in the context, producing types that should be closer to the expectations.
In "Tactic Notation" or "TACTIC EXTEND", entry "constr_with_bindings" now uses type classes and rejects terms with unresolved holes, like entry "constr" does. To get the former behavior use "open_constr_with_bindings" (possible source of incompatibility).
New e-variants eassert, eenough, epose proof, eset, eremember, epose which behave like the corresponding variants with no "e" but turn unresolved implicit arguments into existential variables, on the shelf, rather than failing.
Tactic injection has become more powerful (closes bug #4890) and its documentation has been updated.
New variants of the
first
andsolve
tacticals that do not rely on parsing rules, meant to define tactic notations.Added support for side effects hooks in
cbv
,cbn
andsimpl
. The side effects are provided via a plugin: https://github.com/herbelin/reduction-effects/It is now possible to take hint database names as parameters in a Ltac definition or a Tactic Notation.
New option
Set Ltac Batch Debug
on top ofSet Ltac Debug
for non-interactive Ltac debug output.
Gallina
Now supporting all kinds of binders, including 'pat, in syntax of record fields.
Commands
Goals context can be printed in a more compact way when
Set Printing Compact Contexts
is activated.Unfocused goals can be printed with the
Set Printing Unfocused
option.Print
now shows the types of let-bindings.The compatibility options for printing primitive projections (
Set Printing Primitive Projection Parameters
andSet Printing Primitive Projection Compatibility
) are now off by default.Possibility to unset the printing of notations in a more fine grained fashion than
Unset Printing Notations
is provided without any user-syntax. The goal is that someone creates a plugin to experiment such a user-syntax, to be later integrated in Coq when stabilized.About
now tells if a reference is a coercion.The deprecated
Save
vernacular and its formSave Theorem id
to close proofs have been removed from the syntax. Please useQed
.Search
now sorts results by relevance (the relevance metric is a weighted sum of number of distinct symbols and size of the term).
Standard Library
New file PropExtensionality.v to explicitly work in the axiomatic context of propositional extensionality.
New file SetoidChoice.v axiomatically providing choice over setoids, and, consequently, choice of representatives in equivalence classes. Various proof-theoretic characterizations of choice over setoids in file ChoiceFacts.v.
New lemmas about iff and about orders on positive and Z.
New lemmas on powerRZ.
Strengthened statement of JMeq_eq_dep (closes bug #4912).
The BigN, BigZ, BigZ libraries are no longer part of the Coq standard library, they are now provided by a separate repository https://github.com/coq/bignums The split has been done just after the Int31 library.
IZR (Reals) has been changed to produce a compact representation of integers. As a consequence, IZR is no longer convertible to INR and lemmas such as INR_IZR_INZ should be used instead.
Real constants are now represented using IZR rather than R0 and R1; this might cause rewriting rules to fail to apply to constants.
Added new notation {x & P} for sigT (without a type for x)
Plugins
The Ssreflect plugin is now distributed with Coq. Its documentation has been integrated as a chapter of the reference manual. This chapter is work in progress so feedback is welcome.
The mathematical proof language (also known as declarative mode) was removed.
A new command Extraction TestCompile has been introduced, not meant for the general user but instead for Coq's test-suite.
The extraction plugin is no longer loaded by default. It must be explicitly loaded with [Require Extraction], which is backwards compatible.
The functional induction plugin (which provides the [Function] vernacular) is no longer loaded by default. It must be explicitly loaded with [Require FunInd], which is backwards compatible.
Dependencies
Support for camlp4 has been removed.
Tools
coq_makefile was completely redesigned to improve its maintainability and the extensibility of generated Makefiles, and to make _CoqProject files more palatable to IDEs. Overview:
_CoqProject files contain only Coq specific data (i.e. the list of files, -R options, ...)
coq_makefile translates _CoqProject to Makefile.conf and copies in the desired location a standard Makefile (that reads Makefile.conf)
Makefile extensions can be implemented in a Makefile.local file (read by the main Makefile) by installing a hook in the extension points provided by the standard Makefile
The current version contains code for retro compatibility that prints warnings when a deprecated feature is used. Please upgrade your _CoqProject accordingly.
Additionally, coq_makefile-made Makefiles now support experimental timing targets
pretty-timed
,pretty-timed-before
,pretty-timed-after
,print-pretty-timed-diff
,print-pretty-single-time-diff
,all.timing.diff
, and the variableTIMING=1
(orTIMING=before
orTIMING=after
); see the documentation for more details.
Build Infrastructure
Note that 'make world' does not build the bytecode binaries anymore. For that, you can use 'make byte' (and 'make install-byte' afterwards). Warning: native and byte compilations should not be mixed in the same instance of 'make -j', otherwise both ocamlc and ocamlopt might race for access to the same .cmi files. In short, use "make -j && make -j byte" instead of "make -j world byte".
Universes
Cumulative inductive types. see prefixes "Cumulative", "NonCumulative" for inductive definitions and the option "Set Polymorphic Inductive Cumulativity" in the reference manual.
New syntax
foo@{_}
to instantiate a polymorphic definition with anonymous universes (can also be used withType
).
XML Protocol and internal changes
See dev/doc/changes.txt
Many bugfixes including #1859, #2884, #3613, #3943, #3994, #4250, #4709, #4720, #4824, #4844, #4911, #5026, #5233, #5275, #5315, #5336, #5360, #5390, #5414, #5417, #5420, #5439, #5449, #5475, #5476, #5482, #5501, #5507, #5520, #5523, #5524, #5553, #5577, #5578, #5589, #5597, #5598, #5607, #5618, #5619, #5620, #5641, #5648, #5651, #5671.
Many bugfixes on OS X and Windows (now the test-suite passes on these platforms too).
Many optimizations.
Many documentation improvements.
Details of changes in 8.7+beta2¶
Tools
In CoqIDE, the "Compile Buffer" command takes account of flags in _CoqProject or other project file.
Improvements around some error messages.
Many bug fixes including two important ones:
Bug #5730: CoqIDE becomes unresponsive on file open.
coq_makefile: make sure compile flags for Coq and coq_makefile are in sync (in particular, make sure the
-safe-string
option is used to compile plugins).
Details of changes in 8.7.0¶
OCaml
Users can pass specific flags to the OCaml optimizing compiler by -using the flambda-opts configure-time option.
Beware that compiling Coq with a flambda-enabled compiler is experimental and may require large amounts of RAM and CPU, see INSTALL for more details.
Details of changes in 8.7.1¶
Compatibility with OCaml 4.06.0.
Many bug fixes, documentation improvements, and user message improvements (for details see the 8.7.1 milestone at https://github.com/coq/coq/milestone/10?closed=1).
Details of changes in 8.7.2¶
Fixed a critical bug in the VM handling of universes (#6677). This bug affected all releases since 8.5.
Improved support for building with OCaml 4.06.0 and external num package.
Many other bug fixes, documentation improvements, and user message improvements (for details, see the 8.7.2 milestone at https://github.com/coq/coq/milestone/11?closed=1).
Version 8.6¶
Summary of changes¶
Coq version 8.6 contains the result of refinements, stabilization of 8.5’s features and cleanups of the internals of the system. Over the year of (now time-based) development, about 450 bugs were resolved and over 100 contributions integrated. The main user visible changes are:
A new, faster state-of-the-art universe constraint checker, by Jacques-Henri Jourdan.
In CoqIDE and other asynchronous interfaces, more fine-grained asynchronous processing and error reporting by Enrico Tassi, making Coq capable of recovering from errors and continue processing the document.
More access to the proof engine features from Ltac: goal management primitives, range selectors and a
typeclasses eauto
engine handling multiple goals and multiple successes, by Cyprien Mangin, Matthieu Sozeau and Arnaud Spiwack.Tactic behavior uniformization and specification, generalization of intro-patterns by Hugo Herbelin and others.
A brand new warning system allowing to control warnings, turn them into errors or ignore them selectively by Maxime Dénès, Guillaume Melquiond, Pierre-Marie Pédrot and others.
Irrefutable patterns in abstractions, by Daniel de Rauglaudre.
The ssreflect subterm selection algorithm by Georges Gonthier and Enrico Tassi is now accessible to tactic writers through the ssrmatching plugin.
Integration of LtacProf, a profiler for Ltac by Jason Gross, Paul Steckler, Enrico Tassi and Tobias Tebbi.
Coq 8.6 also comes with a bunch of smaller-scale changes and improvements regarding the different components of the system. We shall only list a few of them.
The iota reduction flag is now a shorthand for match, fix and cofix flags controlling the corresponding reduction rules (by Hugo Herbelin and Maxime Dénès).
Maxime Dénès maintained the native compilation machinery.
Pierre-Marie Pédrot separated the Ltac code from general purpose tactics, and generalized and rationalized the handling of generic arguments, allowing to create new versions of Ltac more easily in the future.
In patterns and terms, @, abbreviations and notations are now interpreted the same way, by Hugo Herbelin.
Name handling for universes has been improved by Pierre-Marie Pédrot and Matthieu Sozeau. The minimization algorithm has been improved by Matthieu Sozeau.
The unifier has been improved by Hugo Herbelin and Matthieu Sozeau, fixing some incompatibilities introduced in Coq 8.5. Unification constraints can now be left floating around and be seen by the user thanks to a new option. The Keyed Unification mode has been improved by Matthieu Sozeau.
The typeclass resolution engine and associated proof search tactic have been reimplemented on top of the proof-engine monad, providing better integration in tactics, and new options have been introduced to control it, by Matthieu Sozeau with help from Théo Zimmermann.
The efficiency of the whole system has been significantly improved thanks to contributions from Pierre-Marie Pédrot, Maxime Dénès and Matthieu Sozeau and performance issue tracking by Jason Gross and Paul Steckler.
Standard library improvements by Jason Gross, Sébastien Hinderer, Pierre Letouzey and others.
Emilio Jesús Gallego Arias contributed many cleanups and refactorings of the pretty-printing and user interface communication components.
Frédéric Besson maintained the micromega tactic.
The OPAM repository for Coq packages has been maintained by Guillaume Claret, Guillaume Melquiond, Matthieu Sozeau, Enrico Tassi and others. A list of packages is now available at https://coq.inria.fr/opam/www/.
Packaging tools and software development kits were prepared by Michael Soegtrop with the help of Maxime Dénès and Enrico Tassi for Windows, and Maxime Dénès and Matthieu Sozeau for MacOS X. Packages are now regularly built on the continuous integration server. Coq now comes with a META file usable with ocamlfind, contributed by Emilio Jesús Gallego Arias, Gregory Malecha, and Matthieu Sozeau.
Matej Košík maintained and greatly improved the continuous integration setup and the testing of Coq contributions. He also contributed many API improvements and code cleanups throughout the system.
The contributors for this version are Bruno Barras, C.J. Bell, Yves Bertot, Frédéric Besson, Pierre Boutillier, Tej Chajed, Guillaume Claret, Xavier Clerc, Pierre Corbineau, Pierre Courtieu, Maxime Dénès, Ricky Elrod, Emilio Jesús Gallego Arias, Jason Gross, Hugo Herbelin, Sébastien Hinderer, Jacques-Henri Jourdan, Matej Košík, Xavier Leroy, Pierre Letouzey, Gregory Malecha, Cyprien Mangin, Erik Martin-Dorel, Guillaume Melquiond, Clément Pit–Claudel, Pierre-Marie Pédrot, Daniel de Rauglaudre, Lionel Rieg, Gabriel Scherer, Thomas Sibut-Pinote, Matthieu Sozeau, Arnaud Spiwack, Paul Steckler, Enrico Tassi, Laurent Théry, Nickolai Zeldovich and Théo Zimmermann. The development process was coordinated by Hugo Herbelin and Matthieu Sozeau with the help of Maxime Dénès, who was also in charge of the release process.
Many power users helped to improve the design of the new features via the bug tracker, the pull request system, the Coq development mailing list or the Coq-Club mailing list. Special thanks to the users who contributed patches and intensive brain-storming and code reviews, starting with Cyril Cohen, Jason Gross, Robbert Krebbers, Jonathan Leivent, Xavier Leroy, Gregory Malecha, Clément Pit–Claudel, Gabriel Scherer and Beta Ziliani. It would however be impossible to mention exhaustively the names of everybody who to some extent influenced the development.
Version 8.6 is the first release of Coq developed on a time-based development cycle. Its development spanned 10 months from the release of Coq 8.5 and was based on a public roadmap. To date, it contains more external contributions than any previous Coq system. Code reviews were systematically done before integration of new features, with an important focus given to compatibility and performance issues, resulting in a hopefully more robust release than Coq 8.5.
Coq Enhancement Proposals (CEPs for short) were introduced by Enrico Tassi to provide more visibility and a discussion period on new features, they are publicly available https://github.com/coq/ceps.
Started during this period, an effort is led by Yves Bertot and Maxime Dénès to put together a Coq consortium.
Potential sources of incompatibilities¶
Symptom: An obligation generated by Program or an abstracted subproof has different arguments.
Cause: Set Shrink Abstract and Set Shrink Obligations are on by default and the subproof does not use the argument.
Remedy:
Adapt the script.
Write an explicit lemma to prove the obligation/subproof and use it instead (compatible with 8.4).
Unset the option for the program/proof the obligation/subproof originates from.
Symptom: In a goal, order of hypotheses, or absence of an equality of the form "x = t" or "t = x", or no unfolding of a local definition.
Cause: This might be connected to a number of fixes in the tactic "subst". The former behavior can be reactivated by issuing "Unset Regular Subst Tactic".
Details of changes in 8.6beta1¶
Kernel
A new, faster state-of-the-art universe constraint checker.
Specification language
Giving implicit arguments explicitly to a constant with multiple choices of implicit arguments does not break any more insertion of further maximal implicit arguments.
Ability to put any pattern in binders, prefixed by quote, e.g. "fun '(a,b) => ...", "λ '(a,(b,c)), ...", "Definition foo '(x,y) := ...". It expands into a "let 'pattern := ..."
Tactics
Flag "Bracketing Last Introduction Pattern" is now on by default.
Flag "Regular Subst Tactic" is now on by default: it respects the initial order of hypothesis, it contracts cycles, it unfolds no local definitions (common source of incompatibilities, fixable by "Unset Regular Subst Tactic").
New flag "Refolding Reduction", now disabled by default, which turns on refolding of constants/fixpoints (as in cbn) during the reductions done during type inference and tactic retyping. Can be extremely expensive. When set off, this recovers the 8.4 behavior of unification and type inference. Potential source of incompatibility with 8.5 developments (the option is set on in Compat/Coq85.v).
New flag "Shrink Abstract" that minimalizes proofs generated by the abstract tactical w.r.t. variables appearing in the body of the proof. On by default and deprecated. Minor source of incompatibility for code relying on the precise arguments of abstracted proofs.
Serious bugs are fixed in tactic "double induction" (source of incompatibilities as soon as the inductive types have dependencies in the type of their constructors; "double induction" remains however deprecated).
In introduction patterns of the form (pat1,...,patn), n should match the exact number of hypotheses introduced (except for local definitions for which pattern can be omitted, as in regular pattern-matching).
Tactic scopes in Ltac like constr: and ltac: now require parentheses around their argument.
Every generic argument type declares a tactic scope of the form "name:(...)" where name is the name of the argument. This generalizes the constr: and ltac: instances.
When in strict mode (i.e. in a Ltac definition), if the "intro" tactic is given a free identifier, it is not bound in subsequent tactics anymore. In order to introduce a binding, use e.g. the "fresh" primitive instead (potential source of incompatibilities).
New tactics is_ind, is_const, is_proj, is_constructor for use in Ltac.
New goal selectors. Sets of goals can be selected by listing integers ranges. Example: "1,4-7,24: tac" focuses "tac" on goals 1,4,5,6,7,24.
For uniformity with "destruct"/"induction" and for a more natural behavior, "injection" can now work in place by activating option "Structural Injection". In this case, hypotheses are also put in the context in the natural left-to-right order and the hypothesis on which injection applies is cleared.
Tactic "contradiction" (hence "easy") now also solve goals with hypotheses of the form "~True" or "t<>t" (possible source of incompatibilities because of more successes in automation, but generally a more intuitive strategy).
Option "Injection On Proofs" was renamed "Keep Proof Equalities". When enabled, injection and inversion do not drop equalities between objects in Prop. Still disabled by default.
New tactics "notypeclasses refine" and "simple notypeclasses refine" that disallow typeclass resolution when typechecking their argument, for use in typeclass hints.
Integration of LtacProf, a profiler for Ltac.
Reduction tactics now accept more fine-grained flags: iota is now a shorthand for the new flags match, fix and cofix.
The ssreflect subterm selection algorithm is now accessible to tactic writers through the ssrmatching plugin.
When used as an argument of an ltac function, "auto" without "with" nor "using" clause now correctly uses only the core hint database by default.
Hints
Revised the syntax of [Hint Cut] to follow standard notation for regexps.
Hint Mode now accepts "!" which means that the mode matches only if the argument's head is not an evar (it goes under applications, casts, and scrutinees of matches and projections).
Hints can now take an optional user-given pattern, used only by [typeclasses eauto] with the [Filtered Unification] option on.
Typeclasses
Many new options and new engine based on the proof monad. The [typeclasses eauto] tactic is now a multi-goal, multi-success tactic. See reference manual for more information. It is planned to replace auto and eauto in the following version. The 8.5 resolution engine is still available to help solve compatibility issues.
Program
The "Shrink Obligations" flag now applies to all obligations, not only those solved by the automatic tactic.
"Shrink Obligations" is on by default and deprecated. Minor source of incompatibility for code relying on the precise arguments of obligations.
Notations
"Bind Scope" can once again bind "Funclass" and "Sortclass".
General infrastructure
New configurable warning system which can be controlled with the vernacular command "Set Warnings", or, under coqc/coqtop, with the flag "-w". In particular, the default is now that warnings are printed by coqc.
In asynchronous mode, Coq is now capable of recovering from errors and continue processing the document.
Tools
coqc accepts a -o option to specify the output file name
coqtop accepts --print-version to print Coq and OCaml versions in easy to parse format
Setting [Printing Dependent Evars Line] can be unset to disable the computation associated with printing the "dependent evars: " line in -emacs mode
Removed the -verbose-compat-notations flag and the corresponding Set Verbose Compat vernacular, since these warnings can now be silenced or turned into errors using "-w".
XML protocol
message format has changed, see dev/doc/changes.txt for more details.
Many bug fixes, minor changes and documentation improvements are not mentioned here.
Details of changes in 8.6¶
Kernel
Fixed critical bug #5248 in VM long multiplication on 32-bit architectures. Was there only since 8.6beta1, so no stable release impacted.
Other bug fixes in universes, type class shelving,...
Details of changes in 8.6.1¶
Fix #5380: Default colors for CoqIDE are actually applied.
Fix plugin warnings
Document named evars (including Show ident)
Fix Bug #5574, document function scope
Adding a test case as requested in bug 5205.
Fix Bug #5568, no dup notation warnings on repeated module imports
Fix documentation of Typeclasses eauto :=
Refactor documentation of records.
Protecting from warnings while compiling 8.6
Fixing an inconsistency between configure and configure.ml
Add test-suite checks for coqchk with constraints
Fix bug #5019 (looping zify on dependent types)
Fix bug 5550: "typeclasses eauto with" does not work with section variables.
Bug 5546, qualify datatype constructors when needed in Show Match
Bug #5535, test for Show with -emacs
Fix bug #5486, don't reverse ids in tuples
Fixing #5522 (anomaly with free vars of pat)
Fix bug #5526, don't check for nonlinearity in notation if printing only
Fix bug #5255
Fix bug #3659: -time should understand multibyte encodings.
FIx bug #5300: Anomaly: Uncaught exception Not_found" in "Print Assumptions".
Fix outdated description in RefMan.
Repairing
Set Rewriting Schemes
Fixing #5487 (v8.5 regression on ltac-matching expressions with evars).
Fix description of command-line arguments for Add (Rec) LoadPath
Fix bug #5377: @? patterns broken.
add XML protocol doc
Fix anomaly when doing [all:Check _.] during a proof.
Correction of bug #4306
Fix #5435: [Eval native_compute in] raises anomaly.
Instances should obey universe binders even when defined by tactics.
Intern names bound in match patterns
funind: Ignore missing info for current function
Do not typecheck twice the type of opaque constants.
show unused intro pattern warning
[future] Be eager when "chaining" already resolved future values.
Opaque side effects
Fix #5132: coq_makefile generates incorrect install goal
Run non-tactic comands without resilient_command
Univs: fix bug #5365, generation of u+k <= v constraints
make
emit
tail recursiveDon't require printing-only notation to be productive
Fix the way setoid_rewrite handles bindings.
Fix for bug 5244 - set printing width ignored when given enough space
Fix bug 4969, autoapply was not tagging shelved subgoals correctly
Version 8.5¶
Summary of changes¶
Coq version 8.5 contains the result of five specific long-term projects:
A new asynchronous evaluation and compilation mode by Enrico Tassi with help from Bruno Barras and Carst Tankink.
Full integration of the new proof engine by Arnaud Spiwack helped by Pierre-Marie Pédrot,
Addition of conversion and reduction based on native compilation by Maxime Dénès and Benjamin Grégoire.
Full universe polymorphism for definitions and inductive types by Matthieu Sozeau.
An implementation of primitive projections with η-conversion bringing significant performance improvements when using records by Matthieu Sozeau.
The full integration of the proof engine, by Arnaud Spiwack and Pierre-Marie Pédrot, brings to primitive tactics and the user level Ltac language dependent subgoals, deep backtracking and multiple goal handling, along with miscellaneous features and an improved potential for future modifications. Dependent subgoals allow statements in a goal to mention the proof of another. Proofs of unsolved subgoals appear as existential variables. Primitive backtracking makes it possible to write a tactic with several possible outcomes which are tried successively when subsequent tactics fail. Primitives are also available to control the backtracking behavior of tactics. Multiple goal handling paves the way for smarter automation tactics. It is currently used for simple goal manipulation such as goal reordering.
The way Coq processes a document in batch and interactive mode has been
redesigned by Enrico Tassi with help from Bruno Barras. Opaque proofs,
the text between Proof and Qed, can be processed asynchronously,
decoupling the checking of definitions and statements from the checking
of proofs. It improves the responsiveness of interactive development,
since proofs can be processed in the background. Similarly, compilation
of a file can be split into two phases: the first one checking only
definitions and statements and the second one checking proofs. A file
resulting from the first phase – with the .vio extension – can be
already Required. All .vio files can be turned into complete .vo files
in parallel. The same infrastructure also allows terminating tactics to
be run in parallel on a set of goals via the par:
goal selector.
CoqIDE was modified to cope with asynchronous checking of the document. Its source code was also made separate from that of Coq, so that CoqIDE no longer has a special status among user interfaces, paving the way for decoupling its release cycle from that of Coq in the future.
Carst Tankink developed a Coq back-end for user interfaces built on Makarius Wenzel’s Prover IDE framework (PIDE), like PIDE/jEdit (with help from Makarius Wenzel) or PIDE/Coqoon (with help from Alexander Faithfull and Jesper Bengtson). The development of such features was funded by the Paral-ITP French ANR project.
The full universe polymorphism extension was designed by Matthieu Sozeau. It conservatively extends the universes system and core calculus with definitions and inductive declarations parameterized by universes and constraints. It is based on a modification of the kernel architecture to handle constraint checking only, leaving the generation of constraints to the refinement/type inference engine. Accordingly, tactics are now fully universe aware, resulting in more localized error messages in case of inconsistencies and allowing higher-level algorithms like unification to be entirely type safe. The internal representation of universes has been modified but this is invisible to the user.
The underlying logic has been extended with η-conversion for records defined with primitive projections by Matthieu Sozeau. This additional form of η-conversion is justified using the same principle than the previously added η-conversion for function types, based on formulations of the Calculus of Inductive Constructions with typed equality. Primitive projections, which do not carry the parameters of the record and are rigid names (not defined as a pattern matching construct), make working with nested records more manageable in terms of time and space consumption. This extension and universe polymorphism were carried out partly while Matthieu Sozeau was working at the IAS in Princeton.
The guard condition has been made compliant with extensional equality
principles such as propositional extensionality and univalence, thanks
to Maxime Dénès and Bruno Barras. To ensure compatibility with the
univalence axiom, a new flag -indices-matter
has been implemented,
taking into account the universe levels of indices when computing the
levels of inductive types. This supports using Coq as a tool to explore
the relations between homotopy theory and type theory.
Maxime Dénès and Benjamin Grégoire developed an implementation of conversion test and normal form computation using the OCaml native compiler. It complements the virtual machine conversion offering much faster computation for expensive functions.
Coq 8.5 also comes with a bunch of many various smaller-scale changes and improvements regarding the different components of the system. We shall only list a few of them.
Pierre Boutillier developed an improved tactic for simplification of
expressions called cbn
.
Maxime Dénès maintained the bytecode-based reduction machine. Pierre Letouzey maintained the extraction mechanism.
Pierre-Marie Pédrot has extended the syntax of terms to, experimentally, allow holes in terms to be solved by a locally specified tactic.
Existential variables are referred to by identifiers rather than mere numbers, thanks to Hugo Herbelin who also improved the tactic language here and there.
Error messages for universe inconsistencies have been improved by Matthieu Sozeau. Error messages for unification and type inference failures have been improved by Hugo Herbelin, Pierre-Marie Pédrot and Arnaud Spiwack.
Pierre Courtieu contributed new features for using Coq through Proof General and for better interactive experience (bullets, Search, etc).
The efficiency of the whole system has been significantly improved thanks to contributions from Pierre-Marie Pédrot.
A distribution channel for Coq packages using the OPAM tool has been initiated by Thomas Braibant and developed by Guillaume Claret, with contributions by Enrico Tassi and feedback from Hugo Herbelin.
Packaging tools were provided by Pierre Letouzey and Enrico Tassi (Windows), Pierre Boutillier, Matthieu Sozeau and Maxime Dénès (MacOS X). Maxime Dénès improved significantly the testing and benchmarking support.
Many power users helped to improve the design of the new features via the bug tracker, the coq development mailing list or the Coq-Club mailing list. Special thanks are going to the users who contributed patches and intensive brain-storming, starting with Jason Gross, Jonathan Leivent, Greg Malecha, Clément Pit-Claudel, Marc Lasson, Lionel Rieg. It would however be impossible to mention with precision all names of people who to some extent influenced the development.
Version 8.5 is one of the most important releases of Coq. Its development spanned over about 3 years and a half with about one year of beta-testing. General maintenance during part or whole of this period has been done by Pierre Boutillier, Pierre Courtieu, Maxime Dénès, Hugo Herbelin, Pierre Letouzey, Guillaume Melquiond, Pierre-Marie Pédrot, Matthieu Sozeau, Arnaud Spiwack, Enrico Tassi as well as Bruno Barras, Yves Bertot, Frédéric Besson, Xavier Clerc, Pierre Corbineau, Jean-Christophe Filliâtre, Julien Forest, Sébastien Hinderer, Assia Mahboubi, Jean-Marc Notin, Yann Régis-Gianas, François Ripault, Carst Tankink. Maxime Dénès coordinated the release process.
Potential sources of incompatibilities¶
List of typical changes to be done to adapt files from Coq 8.4
to Coq 8.5 when not using compatibility option -compat 8.4
.
Symptom: "The reference omega was not found in the current environment".
Cause: "Require Omega" does not import the tactic "omega" any more
Possible solutions:
use "Require Import OmegaTactic" (not compatible with 8.4)
use "Require Import Omega" (compatible with 8.4)
add definition "Ltac omega := Coq.omega.Omega.omega."
Symptom: "intuition" cannot solve a goal (not working anymore on nonstandard connective)
Cause: "intuition" had an accidental non-uniform behavior fixed on nonstandard connectives
Possible solutions:
use "dintuition" instead; it is stronger than "intuition" and works uniformly on nonstandard connectives, such as n-ary conjunctions or disjunctions (not compatible with 8.4)
do the script differently
Symptom: The constructor foo (in type bar) expects n arguments.
Cause: parameters must now be given in patterns
Possible solutions:
use option "Set Asymmetric Patterns" (compatible with 8.4)
add "_" for the parameters (not compatible with 8.4)
turn the parameters into implicit arguments (compatible with 8.4)
Symptom: "NPeano.Nat.foo" not existing anymore
Possible solutions:
use "Nat.foo" instead
Symptom: typing problems with proj1_sig or similar
Cause: coercion from sig to sigT and similar coercions have been removed so as to make the initial state easier to understand for beginners
Solution: change proj1_sig into projT1 and similarly (compatible with 8.4)
Other detailed changes
options for coq compilation (see below for ocaml).
[-I foo] is now deprecated and will not add directory foo to the coq load path (only for ocaml, see below). Just replace [-I foo] by [-Q foo ""] in your project file and re-generate makefile. Or perform the same operation directly in your makefile if you edit it by hand.
Option -R Foo bar is the same in v8.5 than in v8.4 concerning coq load path.
Option [-I foo -as bar] is unchanged but discouraged unless you compile ocaml code. Use -Q foo bar instead.
for more details: see section "Customization at launch time" of the reference manual.
Command line options for ocaml Compilation of ocaml code (plugins)
[-I foo] is not deprecated to add foo to the ocaml load path.
[-I foo -as bar] adds foo to the ocaml load path and adds foo to the coq load path with logical name bar (shortcut for -I foo -Q foo bar).
for more details: section "Customization at launch time" of the reference manual.
Universe Polymorphism.
Refinement, unification and tactics are now aware of universes, resulting in more localized errors. Universe inconsistencies should no more get raised at Qed time but during the proof. Unification always produces well-typed substitutions, hence some rare cases of unifications that succeeded while producing ill-typed terms before will now fail.
The [change p with c] tactic semantics changed, now typechecking [c] at each matching occurrence [t] of the pattern [p], and converting [t] with [c].
Template polymorphic inductive types: the partial application of a template polymorphic type (e.g. list) is not polymorphic. An explicit parameter application (e.g [fun A => list A]) or [apply (list _)] will result in a polymorphic instance.
The type inference algorithm now takes opacity of constants into account. This may have effects on tactics using type inference (e.g. induction). Extra "Transparent" might have to be added to revert opacity of constants.
Type classes.
When writing an
Instance foo : Class A := {| proj := t |}
(note the vertical bars), support for typechecking the projections using the type information and switching to proof mode is no longer available. Use{ }
(without the vertical bars) instead.
Tactic abstract.
Auxiliary lemmas generated by the abstract tactic are removed from the global environment and inlined in the proof term when a proof is ended with Qed. The behavior of 8.4 can be obtained by ending proofs with "Qed exporting" or "Qed exporting ident, .., ident".
Details of changes in 8.5beta1¶
Logic
Primitive projections for records allow for a compact representation of projections, without parameters and avoid the behavior of defined projections that can unfold to a case expression. To turn the use of native projections on, use [Set Primitive Projections]. Record, Class and Structure types defined while this option is set will be defined with primitive projections instead of the usual encoding as a case expression. For compatibility, when p is a primitive projection, @p can be used to refer to the projection with explicit parameters, i.e. [@p] is definitionally equal to [λ params r. r.(p)]. Records with primitive projections have eta-conversion, the canonical form being [mkR pars (p1 t) ... (pn t)].
New universe polymorphism (see reference manual)
New option -type-in-type to collapse the universe hierarchy (this makes the logic inconsistent).
The guard condition for fixpoints is now a bit stricter. Propagation of subterm value through pattern matching is restricted according to the return predicate. Restores compatibility of Coq's logic with the propositional extensionality axiom. May create incompatibilities in recursive programs heavily using dependent types.
Trivial inductive types are no longer defined in Type but in Prop, which leads to a non-dependent induction principle being generated in place of the dependent one. To recover the old behavior, explicitly define your inductive types in Set.
Commands
A command "Variant" allows to define non-recursive variant types.
The command "Record foo ..." does not generate induction principles (foo_rect, foo_rec, foo_ind) anymore by default (feature wish #2693). The command "Variant foo ..." does not either. A flag "Set/Unset Nonrecursive Elimination Schemes" allows changing this. The tactic "induction" on a "Record" or a "Variant" is now actually doing "destruct".
The "Open Scope" command can now be given also a delimiter (e.g. Z).
The "Definition" command now allows the "Local" modifier, allowing for non-importable definitions. The same goes for "Axiom" and "Parameter".
Section-specific commands such as "Let" (resp. "Variable", "Hypothesis") used out of a section now behave like the corresponding "Local" command, i.e. "Local Definition" (resp. "Local Parameter", "Local Axiom"). (potential source of rare incompatibilities).
The "Let" command can now define local (co)fixpoints.
Command "Search" has been renamed into "SearchHead". The command name "Search" now behaves like former "SearchAbout". The latter name is deprecated.
"Search", "About", "SearchHead", "SearchRewrite" and "SearchPattern" now search for hypothesis (of the current goal by default) first. They now also support the goal selector prefix to specify another goal to search: e.g. "n:Search id". This is also true for SearchAbout although it is deprecated.
The coq/user-contrib directory and the XDG directories are no longer recursively added to the load path, so files from installed libraries now need to be fully qualified for the "Require" command to find them. The tools/update-require script can be used to convert a development.
A new Print Strategies command allows visualizing the opacity status of the whole engine.
The "Locate" command now searches through all sorts of qualified namespaces of Coq: terms, modules, tactics, etc. The old behavior of the command can be retrieved using the "Locate Term" command.
New "Derive" command to help writing program by derivation.
New "Refine Instance Mode" option that allows to deactivate the generation of obligations in incomplete typeclass instances, raising an error instead.
"Collection" command to name sets of section hypotheses. Named collections can be used in the syntax of "Proof using" to assert which section variables are used in a proof.
The "Optimize Proof" command can be placed in the middle of a proof to force the compaction of the data structure used to represent the ongoing proof (evar map). This may result in a lower memory footprint and speed up the execution of the following tactics.
"Optimize Heap" command to tell the OCaml runtime to perform a major garbage collection step and heap compaction.
Instance
no longer treats the{|...|}
syntax specially; it handles it in the same way as other commands, e.g. "Definition". Use the{...}
syntax (no pipe symbols) to recover the old behavior.
Specification Language
Slight changes in unification error messages.
Added a syntax $(...)$ that allows putting tactics in terms (may break user notations using "$(", fixable by inserting a space or rewriting the notation).
Constructors in pattern-matching patterns now respect the same rules regarding implicit arguments as in applicative position. The old behavior can be recovered by the command "Set Asymmetric Patterns". As a side effect, notations for constructors explicitly mentioning non-implicit parameters can now be used in patterns. Considering that the pattern language is already rich enough, binding local definitions is however now forbidden in patterns (source of incompatibilities for local definitions that delta-reduce to a constructor).
Type inference algorithm now granting opacity of constants. This might also affect behavior of tactics (source of incompatibilities, solvable by re-declaring transparent constants which were set opaque).
Existential variables are now referred to by an identifier and the relevant part of their instance is displayed by default. They can be reparsed. The naming policy is yet unstable and subject to changes in future releases.
Tactics
New tactic engine allowing dependent subgoals, fully backtracking (also known as multiple success) tactics, as well as tactics which can consider multiple goals together. In the new tactic engine, instantiation information of existential variables is always propagated to tactics, removing the need to manually use the "instantiate" tactics to mark propagation points.
New tactical (a+b) inserts a backtracking point. When (a+b);c fails during the execution of c, it can backtrack and try b instead of a.
New tactical (once a) removes all the backtracking points from a (i.e. it selects the first success of a).
Tactic "constructor" is now fully backtracking. In case of incompatibilities (e.g. combinatoric explosion), the former behavior of "constructor" can be retrieved by using instead "[> once constructor ..]". Thanks to backtracking, undocumented "constructor <tac>" syntax is now equivalent to "[> once (constructor; tac) ..]".
New "multimatch" variant of "match" tactic which backtracks to new branches in case of a later failure. The "match" tactic is equivalent to "once multimatch".
New selector "all:" such that "all:tac" applies tactic "tac" to all the focused goals, instead of just the first one as is the default.
A corresponding new option Set Default Goal Selector "all" makes the tactics in scripts be applied to all the focused goal by default
New selector "par:" such that "par:tac" applies the (terminating) tactic "tac" to all the focused goal in parallel. The number of worker can be selected with -async-proofs-tac-j and also limited using the coqworkmgr utility.
New tactics "revgoals", "cycle" and "swap" to reorder goals.
The semantics of recursive tactics (introduced with "Ltac t := ..." or "let rec t := ... in ...") changed slightly as t is now applied to every goal, not each goal independently. In particular it may be applied when no goals are left. This may cause tactics such as "let rec t := constructor;t" to loop indefinitely. The simple fix is to rewrite the recursive calls as follows: "let rec t := constructor;[t..]" which recovers the earlier behavior (source of rare incompatibilities).
New tactic language feature "numgoals" to count number of goals. It is accompanied by a "guard" tactic which fails if a Boolean test over integers does not pass.
New tactical "[> ... ]" to apply tactics to individual goals.
New tactic "gfail" which works like "fail" except it will also fail if every goal has been solved.
The refine tactic is changed not to use an ad hoc typing algorithm to generate subgoals. It also uses the dependent subgoal feature to generate goals to materialize every existential variable which is introduced by the refinement (source of incompatibilities).
A tactic shelve is introduced to manage the subgoals which may be solved by unification: shelve removes every goal it is applied to from focus. These goals can later be called back into focus by the Unshelve command.
A variant shelve_unifiable only removes those goals which appear as existential variables in other goals. To emulate the old refine, use "refine c;shelve_unifiable". This can still cause incompatibilities in rare occasions.
New "give_up" tactic to skip over a goal. A proof containing given up goals cannot be closed with "Qed", but only with "Admitted".
The implementation of the admit tactic has changed: no axiom is generated for the admitted sub proof. "admit" is now an alias for "give_up". Code relying on this specific behavior of "admit" can be made to work by:
Adding an "Axiom" for each admitted subproof.
Adding a single "Axiom proof_admitted : False." and the Ltac definition "Ltac admit := case proof_admitted.".
Matching using "lazymatch" was fundamentally modified. It now behaves like "match" (immediate execution of the matching branch) but without the backtracking mechanism in case of failure.
New "tryif t then u else v" tactical which executes "u" in case of success of "t" and "v" in case of failure.
New conversion tactic "native_compute": evaluates the goal (or an hypothesis) with a call-by-value strategy, using the OCaml native compiler. Useful on very intensive computations.
New "cbn" tactic, a well-behaved simpl.
Repeated identical calls to omega should now produce identical proof terms.
Tactics btauto, a reflexive Boolean tautology solver.
Tactic "tauto" was exceptionally able to destruct other connectives than the binary connectives "and", "or", "prod", "sum", "iff". This non-uniform behavior has been fixed (bug #2680) and tauto is slightly weaker (possible source of incompatibilities). On the opposite side, new tactic "dtauto" is able to destruct any record-like inductive types, superseding the old version of "tauto".
Similarly, "intuition" has been made more uniform and, where it now fails, "dintuition" can be used (possible source of incompatibilities).
New option "Unset Intuition Negation Unfolding" for deactivating automatic unfolding of "not" in intuition.
Tactic notations can now be defined locally to a module (use "Local" prefix).
Tactic "red" now reduces head beta-iota redexes (potential source of rare incompatibilities).
Tactic "hnf" now reduces inner beta-iota redexes (potential source of rare incompatibilities).
Tactic "intro H" now reduces beta-iota redexes if these hide a product (potential source of rare incompatibilities).
In Ltac matching on patterns of the form "_ pat1 ... patn" now behaves like if matching on "?X pat1 ... patn", i.e. accepting "_" to be instantiated by an applicative term (experimental at this stage, potential source of incompatibilities).
In Ltac matching on goal, types of hypotheses are now interpreted in the %type scope (possible source of incompatibilities).
"change ... in ..." and "simpl ... in ..." now properly consider nested occurrences (possible source of incompatibilities since this alters the numbering of occurrences), but do not support nested occurrences.
Tactics simpl, vm_compute and native_compute can be given a notation string to a constant as argument.
When given a reference as argument, simpl, vm_compute and native_compute now strictly interpret it as the head of a pattern starting with this reference.
The "change p with c" tactic semantics changed, now type checking "c" at each matching occurrence "t" of the pattern "p", and converting "t" with "c".
Now "appcontext" and "context" behave the same. The old buggy behavior of "context" can be retrieved at parse time by setting the "Tactic Compat Context" flag (possible source of incompatibilities).
New introduction pattern p/c which applies lemma c on the fly on the hypothesis under consideration before continuing with introduction pattern p.
New introduction pattern [= x1 .. xn] applies "injection as [x1 .. xn]" on the fly if injection is applicable to the hypothesis under consideration (idea borrowed from Georges Gonthier). Introduction pattern [=] applies "discriminate" if a discriminable equality.
New introduction patterns * and ** to respectively introduce all forthcoming dependent variables and all variables/hypotheses dependent or not.
Tactic "injection c as ipats" now clears c if c refers to an hypothesis and moves the resulting equations in the hypotheses independently of the number of ipats, which has itself to be less than the number of new hypotheses (possible source of incompatibilities; former behavior obtainable by "Unset Injection L2R Pattern Order").
Tactic "injection" now automatically simplifies subgoals "existT n p = existT n p'" into "p = p'" when "n" is in an inductive type for which a decidable equality scheme has been generated with "Scheme Equality" (possible source of incompatibilities).
New tactic "rewrite_strat" for generalized rewriting with user-defined strategies, subsuming autorewrite.
Injection can now also deduce equality of arguments of sort Prop, by using the option "Set Injection On Proofs" (disabled by default). Also improved the error messages.
Tactic "subst id" now supports id occurring in dependent local definitions.
Bugs fixed about intro-pattern "*" might lead to some rare incompatibilities.
New tactical "time" to display time spent executing its argument.
Tactics referring or using a constant dependent in a section variable which has been cleared or renamed in the current goal context now fail (possible source of incompatibilities solvable by avoiding clearing the relevant hypotheses).
New construct "uconstr:c" and "type_term c" to build untyped terms.
Binders in terms defined in Ltac (either "constr" or "uconstr") can now take their names from identifiers defined in Ltac. As a consequence, a name cannot be used in a binder "constr:(fun x => ...)" if an Ltac variable of that name already exists and does not contain an identifier. Source of occasional incompatibilities.
The "refine" tactic now accepts untyped terms built with "uconstr" so that terms with holes can be constructed piecewise in Ltac.
New bullets --, ++, , ---, +++, *, ... made available.
More informative messages when wrong bullet is used.
Bullet suggestion when a subgoal is solved.
New tactic "enough", symmetric to "assert", but with subgoals swapped, as a more friendly replacement of "cut".
In destruct/induction, experimental modifier "!" prefixing the hypothesis name to tell not erasing the hypothesis.
Bug fixes in "inversion as" may occasionally lead to incompatibilities.
Behavior of introduction patterns -> and <- made more uniform (hypothesis is cleared, rewrite in hypotheses and conclusion and erasing the variable when rewriting a variable).
New experimental option "Set Standard Proposition Elimination Names" so that case analysis or induction on schemes in Type containing propositions now produces "H"-based names.
Tactics from plugins are now active only when the corresponding module is imported (source of incompatibilities, solvable by adding an "Import"; in the particular case of Omega, use "Require Import OmegaTactic").
Semantics of destruct/induction has been made more regular in some edge cases, possibly leading to incompatibilities:
new goals are now opened when the term does not match a subterm of the goal and has unresolved holes, while in 8.4 these holes were turned into existential variables
when no "at" option is given, the historical semantics which selects all subterms syntactically identical to the first subterm matching the given pattern is used
non-dependent destruct/induction on an hypothesis with premises in an inductive type with indices is fixed
residual local definitions are now correctly removed.
The rename tactic may now replace variables in parallel.
A new "Info" command replaces the "info" tactical discontinued in v8.4. It still gives informative results in many cases.
The "info_auto" tactic is known to be broken and does not print a trace anymore. Use "Info 1 auto" instead. The same goes for "info_trivial". On the other hand "info_eauto" still works fine, while "Info 1 eauto" prints a trivial trace.
When using a lemma of the prototypical form "forall A, {a:A & P a}", "apply" and "apply in" do not instantiate anymore "A" with the current goal and use "a" as the proof, as they were sometimes doing, now considering that it is a too powerful decision.
Program
"Solve Obligations using" changed to "Solve Obligations with", consistent with "Proof with".
Program Lemma, Definition now respect automatic introduction.
Program Lemma, Definition, etc.. now interpret "->" like Lemma and Definition as a non-dependent arrow (potential source of incompatibility).
Add/document "Set Hide Obligations" (to hide obligations in the final term inside an implicit argument) and "Set Shrink Obligations" (to minimize dependencies of obligations defined by tactics).
Notations
The syntax "x -> y" is now declared at level 99. In particular, it has now a lower priority than "<->": "A -> B <-> C" is now "A -> (B <-> C)" (possible source of incompatibilities)
Notations accept term-providing tactics using the $(...)$ syntax.
"Bind Scope" can no longer bind "Funclass" and "Sortclass".
A notation can be given a (compat "8.x") annotation, making it behave like a "only parsing" notation, but the annotation may lead to eventually issue warnings or errors in further versions when this notation is used.
More systematic insertion of spaces as a default for printing notations ("format" still available to override the default).
In notations, a level modifier referring to a non-existent variable is now considered an error rather than silently ignored.
Tools
Option -I now only adds directories to the ml path.
Option -Q behaves as -R, except that the logical path of any loaded file has to be fully qualified.
Option -R no longer adds recursively to the ml path; only the root directory is added. (Behavior with respect to the load path is unchanged.)
Option -nois prevents coq/theories and coq/plugins to be recursively added to the load path. (Same behavior as with coq/user-contrib.)
coqdep accepts a -dumpgraph option generating a dot file.
Makefiles generated through coq_makefile have three new targets "quick" "checkproofs" and "vio2vo", allowing respectively to asynchronously compile the files without playing the proof scripts, asynchronously checking that the quickly generated proofs are correct and generating the object files from the quickly generated proofs.
The XML plugin was discontinued and removed from the source.
A new utility called coqworkmgr can be used to limit the number of concurrent workers started by independent processes, like make and CoqIDE. This is of interest for users of the par: goal selector.
Interfaces
CoqIDE supports asynchronous edition of the document, ongoing tasks and errors are reported in the bottom right window. The number of workers taking care of processing proofs can be selected with -async-proofs-j.
CoqIDE highlights in yellow "unsafe" commands such as axiom declarations, and tactics like "give_up".
CoqIDE supports Proof General like key bindings; to activate the PG mode go to Edit -> Preferences -> Editor. For the documentation see Help -> Help for PG mode.
CoqIDE automatically retracts the locked area when one edits the locked text.
CoqIDE search and replace got regular expressions power. See the documentation of OCaml's Str module for the supported syntax.
Many CoqIDE windows, including the query one, are now detachable to improve usability on multi screen work stations.
Coqtop/coqc outputs highlighted syntax. Colors can be configured thanks to the COQ_COLORS environment variable, and their current state can be displayed with the -list-tags command line option.
Third party user interfaces can install their main loop in $COQLIB/toploop and call coqtop with the -toploop flag to select it.
Internal Infrastructure
Many reorganizations in the ocaml source files. For instance, many internal a.s.t. of Coq are now placed in mli files in a new directory intf/, for instance constrexpr.mli or glob_term.mli. More details in dev/doc/changes.
The file states/initial.coq does not exist anymore. Instead, coqtop initially does a "Require" of Prelude.vo (or nothing when given the options -noinit or -nois).
The format of vo files has slightly changed: cf final comments in checker/cic.mli.
The build system does not produce anymore programs named coqtop.opt and a symbolic link to coqtop. Instead, coqtop is now directly an executable compiled with the best OCaml compiler available. The bytecode program coqtop.byte is still produced. Same for other utilities.
Some options of the ./configure script slightly changed:
The -coqrunbyteflags and its blank-separated argument is replaced by option -vmbyteflags which expects a comma-separated argument.
The -coqtoolsbyteflags option is discontinued, see -no-custom instead.
Miscellaneous
ML plugins now require a "DECLARE PLUGIN "foo"" statement. The "foo" name must be exactly the name of the ML module that will be loaded through a "Declare ML "foo"" command.
Details of changes in 8.5beta2¶
Logic
The VM now supports inductive types with up to 8388851 non-constant constructors and up to 8388607 constant ones.
Specification language
Syntax "$(tactic)$" changed to "ltac: tactic".
Tactics
A script using the admit tactic can no longer be concluded by either Qed or Defined. In the first case, Admitted can be used instead. In the second case, a subproof should be used.
The easy tactic and the now tactical now have a more predictable behavior, but they might now discharge some previously unsolved goals.
Extraction
Definitions extracted to Haskell GHC should no longer randomly segfault when some Coq types cannot be represented by Haskell types.
Definitions can now be extracted to Json for post-processing.
Tools
Option -I -as has been removed, and option -R -as has been deprecated. In both cases, option -R can be used instead.
coq_makefile now generates double-colon rules for rules such as clean.
API
The interface of [change] has changed to take a [change_arg], which can be built from a [constr] using [make_change_arg].
Details of changes in 8.5beta3¶
Commands
New command "Redirect" to redirect the output of a command to a file.
New command "Undelimit Scope" to remove the delimiter of a scope.
New option "Strict Universe Declaration", set by default. It enforces the declaration of all polymorphic universes appearing in a definition when introducing it.
New command "Show id" to show goal named id.
Option "Virtual Machine" removed.
Tactics
New flag "Regular Subst Tactic" which fixes "subst" in situations where it failed to substitute all substitutable equations or failed to simplify cycles, or accidentally unfolded local definitions (flag is off by default).
New flag "Loose Hint Behavior" to handle hints loaded but not imported in a special way. It accepts three distinct flags: * "Lax", which is the default one, sets the old behavior, i.e. a non-imported hint behaves the same as an imported one. * "Warn" outputs a warning when a non-imported hint is used. Note that this is an over-approximation, because a hint may be triggered by an eauto run that will eventually fail and backtrack. * "Strict" changes the behavior of an unloaded hint to the one of the fail tactic, allowing to emulate the hopefully future import-scoped hint mechanism.
New compatibility flag "Universal Lemma Under Conjunction" which let tactics working under conjunctions apply sublemmas of the form "forall A, ... -> A".
New compatibility flag "Bracketing Last Introduction Pattern" which can be set so that the last disjunctive-conjunctive introduction pattern given to "intros" automatically complete the introduction of its subcomponents, as the the disjunctive-conjunctive introduction patterns in non-terminal position already do.
New flag "Shrink Abstract" that minimalizes proofs generated by the abstract tactical w.r.t. variables appearing in the body of the proof.
Program
The "Shrink Obligations" flag now applies to all obligations, not only those solved by the automatic tactic.
Importing Program no longer overrides the "exists" tactic (potential source of incompatibilities).
Hints costs are now correctly taken into account (potential source of incompatibilities).
Documented the Hint Cut command that allows control of the proof search during typeclass resolution (see reference manual).
API
Some functions from pretyping/typing.ml and their derivatives were potential source of evarmap leaks, as they dropped their resulting evarmap. The situation was clarified by renaming them according to a
unsafe_*
scheme. Their sound variant is likewise renamed to their old name. The following renamings were made.Typing.type_of
->unsafe_type_of
Typing.e_type_of
->type_of
A new
e_type_of
function that matches thee_
prefix policyTacmach.pf_type_of
->pf_unsafe_type_of
A new safe
pf_type_of
function.
All uses of
unsafe_*
functions should be eventually eliminated.
Tools
Added an option -w to control the output of coqtop warnings.
Configure now takes an optional -native-compiler (yes|no) flag replacing -no-native-compiler. The new flag is set to no by default under Windows.
Flag -no-native-compiler was removed and became the default for coqc. If precompilation of files for native conversion test is desired, use -native-compiler.
The -compile command-line option now takes the full path of the considered file, including the ".v" extension, and outputs a warning if such an extension is lacking.
The -require and -load-vernac-object command-line options now take a logical path of a given library rather than a physical path, thus they behave like Require [Import] path.
The -vm command-line option has been removed.
Standard Library
There is now a Coq.Compat.Coq84 library, which sets the various compatibility options and does a few redefinitions to make Coq behave more like Coq v8.4. The standard way of putting Coq in v8.4 compatibility mode is to pass the command line flags "-require Coq.Compat.Coq84 -compat 8.4".
Details of changes in 8.5¶
Tools
Flag "-compat 8.4" now loads Coq.Compat.Coq84. The standard way of putting Coq in v8.4 compatibility mode is to pass the command line flag "-compat 8.4". It can be followed by "-require Coq.Compat.AdmitAxiom" if the 8.4 behavior of admit is needed, in which case it uses an axiom.
Specification language
Syntax "$(tactic)$" changed to "ltac:(tactic)".
Tactics
Syntax "destruct !hyp" changed to "destruct (hyp)", and similarly for induction (rare source of incompatibilities easily solvable by removing parentheses around "hyp" when not for the purpose of keeping the hypothesis).
Syntax "p/c" for on-the-fly application of a lemma c before introducing along pattern p changed to p%c1..%cn. The feature and syntax are in experimental stage.
"Proof using" does not clear unused section variables.
Tactic "refine" has been changed back to the 8.4 behavior of shelving subgoals that occur in other subgoals. The "refine" tactic of 8.5beta3 has been renamed "simple refine"; it does not shelve any subgoal.
New tactical "unshelve tac" which grab existential variables put on the tactic shelve by the execution of "tac".
Details of changes in 8.5pl1¶
Critical bugfix
The subterm relation for the guard condition was incorrectly defined on primitive projections (#4588)
Plugin development tools
add a .merlin target to the makefile
Various performance improvements (time, space used by .vo files)
Other bugfixes
Fix order of arguments to Big.compare_case in ExtrOcamlZBigInt.v
Added compatibility coercions from Specif.v which were present in Coq 8.4.
Fixing a source of inefficiency and an artificial dependency in the printer in the congruence tactic.
Allow to unset the refinement mode of Instance in ML
Fixing an incorrect use of prod_appvect on a term which was not a product in setoid_rewrite.
Add -compat 8.4 econstructor tactics, and tests
Add compatibility Nonrecursive Elimination Schemes
Fixing the "No applicable tactic" uninformative error message regression on apply.
Univs: fix get_current_context (bug #4603, part I)
Fix a bug in Program coercion code
Fix handling of arity of definitional classes.
#4630: Some tactics are 20x slower in 8.5 than 8.4.
#4627: records with no declared arity can be template polymorphic.
#4623: set tactic too weak with universes (regression)
Fix incorrect behavior of CS resolution
#4591: Uncaught exception in directory browsing.
CoqIDE is more resilient to initialization errors.
#4614: "Fully check the document" is uninterruptible.
Try eta-expansion of records only on non-recursive ones
Fix bug when a sort is ascribed to a Record
Primitive projections: protect kernel from erroneous definitions.
Fixed bug #4533 with previous Keyed Unification commit
Win: kill unreliable hence do not waitpid after kill -9 (Close #4369)
Fix strategy of Keyed Unification
#4608: Anomaly "output_value: abstract value (outside heap)".
#4607: do not read native code files if native compiler was disabled.
#4105: poor escaping in the protocol between CoqIDE and coqtop.
#4596: [rewrite] broke in the past few weeks.
#4533 (partial): respect declared global transparency of projections in unification.ml
#4544: Backtrack on using full betaiota reduction during keyed unification.
#4540: CoqIDE bottom progress bar does not update.
Fix regression from 8.4 in reflexivity
#4580: [Set Refine Instance Mode] also used for Program Instance.
#4582: cannot override notation [ x ]. MAY CREATE INCOMPATIBILITIES, see #4683.
STM: Print/Extraction have to be skipped if -quick
#4542: CoqIDE: STOP button also stops workers
STM: classify some variants of Instance as regular ``
Fork `
nodes.#4574: Anomaly: Uncaught exception Invalid_argument("splay_arity").
Do not give a name to anonymous evars anymore. See bug #4547.
STM: always stock in vio files the first node (state) of a proof
STM: not delegate proofs that contain Vernac(Module|Require|Import), #4530
Don't fail fatally if PATH is not set.
#4537: Coq 8.5 is slower in typeclass resolution.
#4522: Incorrect "Warning..." on windows.
#4373: coqdep does not know about .vio files.
#3826: "Incompatible module types" is uninformative.
#4495: Failed assertion in metasyntax.ml.
#4511: evar tactic can create non-typed evars.
#4503: mixing universe polymorphic and monomorphic variables and definitions in sections is unsupported.
#4519: oops, global shadowed local universe level bindings.
#4506: Anomaly: File "pretyping/indrec.ml", line 169, characters 14-20: Assertion failed.
#4548: CoqIDE crashes when going back one command
Details of changes in 8.5pl2¶
Critical bugfix
Checksums of .vo files dependencies were not correctly checked.
Unicode-to-ASCII translation was not injective, leading in a soundness bug in the native compiler.
Other bugfixes
#4097: more efficient occur-check in presence of primitive projections
#4398: type_scope used consistently in "match goal".
#4450: eauto does not work with polymorphic lemmas
#4677: fix alpha-conversion in notations needing eta-expansion.
Fully preserve initial order of hypotheses in "Regular Subst Tactic" mode.
#4644: a regression in unification.
#4725: Function (Error: Conversion test raised an anomaly) and Program (Error: Cannot infer this placeholder of type)
#4747: Problem building Coq 8.5pl1 with OCaml 4.03.0: Fatal warnings
#4752: CoqIDE crash on files not ended by ".v".
#4777: printing inefficiency with implicit arguments
#4818: "Admitted" fails due to undefined universe anomaly after calling "destruct"
#4823: remote counter: avoid thread race on sockets
#4841: -verbose flag changed semantics in 8.5, is much harder to use
#4851: [nsatz] cannot handle duplicated hypotheses
#4858: Anomaly: Uncaught exception Failure("hd"). Please report. in variant of nsatz
#4880: [nsatz_compute] generates invalid certificates if given redundant hypotheses
#4881: synchronizing "Declare Implicit Tactic" with backtrack.
#4882: anomaly with Declare Implicit Tactic on hole of type with evars
Fix use of "Declare Implicit Tactic" in refine. triggered by CoqIDE
#4069, #4718: congruence fails when universes are involved.
Universes
Disallow silently dropping universe instances applied to variables (forward compatible)
Allow explicit universe instances on notations, when they can apply to the head reference of their expansion.
Build infrastructure
New update on how to find camlp5 binary and library at configure time.
Details of changes in 8.5pl3¶
Critical bugfix
#4876: Guard checker incompleteness when using primitive projections
Other bugfixes
#4780: Induction with universe polymorphism on was creating ill-typed terms.
#4673: regression in setoid_rewrite, unfolding let-ins for type unification.
#4754: Regression in setoid_rewrite, allow postponed unification problems to remain.
#4769: Anomaly with universe polymorphic schemes defined inside sections.
#3886: Program: duplicate obligations of mutual fixpoints.
#4994: Documentation typo.
#5008: Use the "md5" command on OpenBSD.
#5007: Do not assume the "TERM" environment variable is always set.
#4606: Output a break before a list only if there was an empty line.
#5001: metas not cleaned properly in clenv_refine_in.
#2336: incorrect glob data for module symbols (bug #2336).
#4832: Remove extraneous dot in error message.
Anomaly in printing a unification error message.
#4947: Options which take string arguments are not backwards compatible.
#4156: micromega cache files are now hidden files.
#4871: interrupting par:abstract kills coqtop.
#5043: [Admitted] lemmas pick up section variables.
Fix name of internal refine ("simple refine").
#5062: probably a typo in Strict Proofs mode.
#5065: Anomaly: Not a proof by induction.
Restore native compiler optimizations, they were disabled since 8.5!
#5077: failure on typing a fixpoint with evars in its type.
Fix recursive notation bug.
#5095: irrelevant too strict test in let-in abstraction.
Ensuring that the evar name is preserved by "rename".
#4887: confusion between using and with in documentation of firstorder.
Bug in subst with let-ins.
#4762: eauto weaker than auto.
Remove if_then_else (was buggy). Use tryif instead.
#4970: confusion between special "{" and non-special "{{" in notations.
#4529: primitive projections unfolding.
#4416: Incorrect "Error: Incorrect number of goals".
#4863: abstract in typeclass hint fails.
#5123: unshelve can impact typeclass resolution
Fix a collision about the meta-variable ".." in recursive notations.
Fix printing of info_auto.
#3209: Not_found due to an occur-check cycle.
#5097: status of evars refined by "clear" in ltac: closed wrt evars.
#5150: Missing dependency of the test-suite subsystems in prerequisite.
Fix a bug in error printing of unif constraints
#3941: Do not stop propagation of signals when Coq is busy.
#4822: Incorrect assertion in cbn.
#3479 parsing of "{" and "}" when a keyword starts with "{" or "}".
#5127: Memory corruption with the VM.
#5102: bullets parsing broken by calls to parse_entry.
Various documentation improvements
Version 8.4¶
Summary of changes¶
Coq version 8.4 contains the result of three long-term projects: a new modular library of arithmetic by Pierre Letouzey, a new proof engine by Arnaud Spiwack and a new communication protocol for CoqIDE by Vincent Gross.
The new modular library of arithmetic extends, generalizes and unifies the existing libraries on Peano arithmetic (types nat, N and BigN), positive arithmetic (type positive), integer arithmetic (Z and BigZ) and machine word arithmetic (type Int31). It provides with unified notations (e.g. systematic use of add and mul for denoting the addition and multiplication operators), systematic and generic development of operators and properties of these operators for all the types mentioned above, including gcd, pcm, power, square root, base 2 logarithm, division, modulo, bitwise operations, logical shifts, comparisons, iterators, ...
The most visible feature of the new proof engine is the support for structured scripts (bullets and proof brackets) but, even if yet not user-available, the new engine also provides the basis for refining existential variables using tactics, for applying tactics to several goals simultaneously, for reordering goals, all features which are planned for the next release. The new proof engine forced Pierre Letouzey to reimplement info and Show Script differently.
Before version 8.4, CoqIDE was linked to Coq with the graphical interface living in a separate thread. From version 8.4, CoqIDE is a separate process communicating with Coq through a textual channel. This allows for a more robust interfacing, the ability to interrupt Coq without interrupting the interface, and the ability to manage several sessions in parallel. Relying on the infrastructure work made by Vincent Gross, Pierre Letouzey, Pierre Boutillier and Pierre-Marie Pédrot contributed many various refinements of CoqIDE.
Coq 8.4 also comes with a bunch of various smaller-scale changes and improvements regarding the different components of the system.
The underlying logic has been extended with η-conversion thanks to Hugo Herbelin, Stéphane Glondu and Benjamin Grégoire. The addition of η-conversion is justified by the confidence that the formulation of the Calculus of Inductive Constructions based on typed equality (such as the one considered in Lee and Werner to build a set-theoretic model of CIC [LW11]) is applicable to the concrete implementation of Coq.
The underlying logic benefited also from a refinement of the guard condition for fixpoints by Pierre Boutillier, the point being that it is safe to propagate the information about structurally smaller arguments through β-redexes that are blocked by the “match” construction (blocked commutative cuts).
Relying on the added permissiveness of the guard condition, Hugo Herbelin could extend the pattern matching compilation algorithm so that matching over a sequence of terms involving dependencies of a term or of the indices of the type of a term in the type of other terms is systematically supported.
Regarding the high-level specification language, Pierre Boutillier
introduced the ability to give implicit arguments to anonymous
functions, Hugo Herbelin introduced the ability to define notations with
several binders (e.g. exists x y z, P
), Matthieu Sozeau made the
typeclass inference mechanism more robust and predictable, Enrico
Tassi introduced a command Arguments that generalizes Implicit Arguments
and Arguments Scope for assigning various properties to arguments of
constants. Various improvements in the type inference algorithm were
provided by Matthieu Sozeau and Hugo Herbelin with contributions from
Enrico Tassi.
Regarding tactics, Hugo Herbelin introduced support for referring to
expressions occurring in the goal by pattern in tactics such as set or
destruct. Hugo Herbelin also relied on ideas from Chung-Kil Hur’s Heq
plugin to introduce automatic computation of occurrences to generalize
when using destruct and induction on types with indices. Stéphane Glondu
introduced new tactics constr_eq
, is_evar
, and has_evar
, to be used
when writing complex tactics. Enrico Tassi added support to fine-tuning
the behavior of simpl
. Enrico Tassi added the ability to specify over
which variables of a section a lemma has to be exactly generalized.
Pierre Letouzey added a tactic timeout and the interruptibility of
vm_compute
. Bug fixes and miscellaneous improvements of the tactic
language came from Hugo Herbelin, Pierre Letouzey and Matthieu Sozeau.
Regarding decision tactics, Loïc Pottier maintained nsatz, moving in particular to a typeclass based reification of goals while Frédéric Besson maintained Micromega, adding in particular support for division.
Regarding commands, Stéphane Glondu provided new commands to analyze the structure of type universes.
Regarding libraries, a new library about lists of a given length (called vectors) has been provided by Pierre Boutillier. A new instance of finite sets based on Red-Black trees and provided by Andrew Appel has been adapted for the standard library by Pierre Letouzey. In the library of real analysis, Yves Bertot changed the definition of \(\pi\) and provided a proof of the long-standing fact yet remaining unproved in this library, namely that \(sin \frac{\pi}{2} = 1\).
Pierre Corbineau maintained the Mathematical Proof Language (C-zar).
Bruno Barras and Benjamin Grégoire maintained the call-by-value reduction machines.
The extraction mechanism benefited from several improvements provided by Pierre Letouzey.
Pierre Letouzey maintained the module system, with contributions from Élie Soubiran.
Julien Forest maintained the Function command.
Matthieu Sozeau maintained the setoid rewriting mechanism.
Coq related tools have been upgraded too. In particular, coq_makefile has been largely revised by Pierre Boutillier. Also, patches from Adam Chlipala for coqdoc have been integrated by Pierre Boutillier.
Bruno Barras and Pierre Letouzey maintained the coqchk
checker.
Pierre Courtieu and Arnaud Spiwack contributed new features for using Coq through Proof General.
The Dp plugin has been removed. Use the plugin provided with Why 3 instead (http://why3.lri.fr/).
Under the hood, the Coq architecture benefited from improvements in terms of efficiency and robustness, especially regarding universes management and existential variables management, thanks to Pierre Letouzey and Yann Régis-Gianas with contributions from Stéphane Glondu and Matthias Puech. The build system is maintained by Pierre Letouzey with contributions from Stéphane Glondu and Pierre Boutillier.
A new backtracking mechanism simplifying the task of external interfaces has been designed by Pierre Letouzey.
The general maintenance was done by Pierre Letouzey, Hugo Herbelin, Pierre Boutillier, Matthieu Sozeau and Stéphane Glondu with also specific contributions from Guillaume Melquiond, Julien Narboux and Pierre-Marie Pédrot.
Packaging tools were provided by Pierre Letouzey (Windows), Pierre Boutillier (MacOS), Stéphane Glondu (Debian). Releasing, testing and benchmarking support was provided by Jean-Marc Notin.
Many suggestions for improvements were motivated by feedback from users, on either the bug tracker or the Coq-Club mailing list. Special thanks are going to the users who contributed patches, starting with Tom Prince. Other patch contributors include Cédric Auger, David Baelde, Dan Grayson, Paolo Herms, Robbert Krebbers, Marc Lasson, Hendrik Tews and Eelis van der Weegen.
Potential sources of incompatibilities¶
The main known incompatibilities between 8.3 and 8.4 are consequences of the following changes:
The reorganization of the library of numbers:
Several definitions have new names or are defined in modules of different names, but a special care has been taken to have this renaming transparent for the user thanks to compatibility notations.
However some definitions have changed, what might require some adaptations. The most noticeable examples are:
The "?=" notation which now bind to Pos.compare rather than former Pcompare (now Pos.compare_cont).
Changes in names may induce different automatically generated names in proof scripts (e.g. when issuing "destruct Z_le_gt_dec").
Z.add has a new definition, hence, applying "simpl" on subterms of its body might give different results than before.
BigN.shiftl and BigN.shiftr have reversed arguments order, the power function in BigN now takes two BigN.
Other changes in libraries:
The definition of functions over "vectors" (list of fixed length) have changed.
TheoryList.v has been removed.
Slight changes in tactics:
Less unfolding of fixpoints when applying destruct or inversion on a fixpoint hiding an inductive type (add an extra call to simpl to preserve compatibility).
Less unexpected local definitions when applying "destruct" (incompatibilities solvable by adapting name hypotheses).
Tactic "apply" might succeed more often, e.g. by now solving pattern-matching of the form ?f x y = g(x,y) (compatibility ensured by using "Unset Tactic Pattern Unification"), but also because it supports (full) betaiota (using "simple apply" might then help).
Tactic autorewrite does no longer instantiate pre-existing existential variables.
Tactic "info" is now available only for auto, eauto and trivial.
Miscellaneous changes:
The command "Load" is now atomic for backtracking (use "Unset Atomic Load" for compatibility).
Details of changes in 8.4beta¶
Logic
Standard eta-conversion now supported (dependent product only).
Guard condition improvement: subterm property is propagated through beta-redex blocked by pattern-matching, as in "(match v with C .. => fun x => u end) x"; this allows for instance to use "rewrite ... in ..." without breaking the guard condition.
Specification language and notations
Maximal implicit arguments can now be set locally by { }. The registration traverses fixpoints and lambdas. Because there is conversion in types, maximal implicit arguments are not taken into account in partial applications (use eta expanded form with explicit { } instead).
Added support for recursive notations with binders (allows for instance to write "exists x y z, P").
Structure/Record printing can be disable by "Unset Printing Records". In addition, it can be controlled on type by type basis using "Add Printing Record" or "Add Printing Constructor".
Pattern-matching compilation algorithm: in "match x, y with ... end", possible dependencies of x (or of the indices of its type) in the type of y are now taken into account.
Tactics
New proof engine.
Scripts can now be structured thanks to bullets - * + and to subgoal delimitation via { }. Note: for use with Proof General, a cvs version of Proof General no older than mid-July 2011 is currently required.
Support for tactical "info" is suspended.
Support for command "Show Script" is suspended.
New tactics constr_eq, is_evar and has_evar for use in Ltac (DOC TODO).
Removed the two-argument variant of "decide equality".
New experimental tactical "timeout <n> <tac>". Since <n> is a time in second for the moment, this feature should rather be avoided in scripts meant to be machine-independent.
Fix in "destruct": removal of unexpected local definitions in context might result in some rare incompatibilities (solvable by adapting name hypotheses).
Introduction pattern "_" made more robust.
Tactic (and Eval command) vm_compute can now be interrupted via Ctrl-C.
Unification in "apply" supports unification of patterns of the form ?f x y = g(x,y) (compatibility ensured by using "Unset Tactic Pattern Unification"). It also supports (full) betaiota.
Tactic autorewrite does no longer instantiate pre-existing existential variables (theoretical source of possible incompatibilities).
Tactic "dependent rewrite" now supports equality in "sig".
Tactic omega now understands Zpred (wish #1912) and can prove any goal from a context containing an arithmetical contradiction (wish #2236).
Using "auto with nocore" disables the use of the "core" database (wish #2188). This pseudo-database "nocore" can also be used with trivial and eauto.
Tactics "set", "destruct" and "induction" accepts incomplete terms and use the goal to complete the pattern assuming it is unambiguous.
When used on arguments with a dependent type, tactics such as "destruct", "induction", "case", "elim", etc. now try to abstract automatically the dependencies over the arguments of the types (based on initial ideas from Chung-Kil Hur, extension to nested dependencies suggested by Dan Grayson)
Tactic "injection" now failing on an equality showing no constructors while it was formerly generalizing again the goal over the given equality.
In Ltac, the "context [...]" syntax has now a variant "appcontext [...]" allowing to match partial applications in larger applications.
When applying destruct or inversion on a fixpoint hiding an inductive type, recursive calls to the fixpoint now remain folded by default (rare source of incompatibility generally solvable by adding a call to simpl).
In an ltac pattern containing a "match", a final "| _ => _" branch could be used now instead of enumerating all remaining constructors. Moreover, the pattern "match _ with _ => _ end" now allows to match any "match". A "in" annotation can also be added to restrict to a precise inductive type.
The behavior of "simpl" can be tuned using the "Arguments" vernacular. In particular constants can be marked so that they are always/never unfolded by "simpl", or unfolded only when a set of arguments evaluates to a constructor. Last one can mark a constant so that it is unfolded only if the simplified term does not expose a match in head position.
Commands
It is now mandatory to have a space (or tabulation or newline or end-of-file) after a "." ending a sentence.
In SearchAbout, the [ ] delimiters are now optional.
New command "Add/Remove Search Blacklist <substring> ...": a Search or SearchAbout or similar query will never mention lemmas whose qualified names contain any of the declared substrings. The default blacklisted substrings are
_subproof
,Private_
.When the output file of "Print Universes" ends in ".dot" or ".gv", the universe graph is printed in the DOT language, and can be processed by Graphviz tools.
New command "Print Sorted Universes".
The undocumented and obsolete option "Set/Unset Boxed Definitions" has been removed, as well as syntaxes like "Boxed Fixpoint foo".
A new option "Set Default Timeout n / Unset Default Timeout".
Qed now uses information from the reduction tactics used in proof script to avoid conversion at Qed time to go into a very long computation.
New command "Show Goal ident" to display the statement of a goal, even a closed one (available from Proof General).
Command "Proof" accept a new modifier "using" to force generalization over a given list of section variables at section ending (DOC TODO).
New command "Arguments" generalizing "Implicit Arguments" and "Arguments Scope" and that also allows to rename the parameters of a definition and to tune the behavior of the tactic "simpl".
Module System
During subtyping checks, an opaque constant in a module type could now be implemented by anything of the right type, even if bodies differ. Said otherwise, with respect to subtyping, an opaque constant behaves just as a parameter. Coqchk was already implementing this, but not coqtop.
The inlining done during application of functors can now be controlled more precisely, by the annotations (no inline) or (inline at level XX). With the latter annotation, only functor parameters whose levels are lower or equal than XX will be inlined. The level of a parameter can be fixed by "Parameter Inline(30) foo". When levels aren't given, the default value is 100. One can also use the flag "Set Inline Level ..." to set a level (DOC TODO).
Print Assumptions should now handle correctly opaque modules (#2168).
Print Module (Type) now tries to print more details, such as types and bodies of the module elements. Note that Print Module Type could be used on a module to display only its interface. The option "Set Short Module Printing" could be used to switch back to the earlier behavior were only field names were displayed.
Libraries
Extension of the abstract part of Numbers, which now provide axiomatizations and results about many more integer functions, such as pow, gcd, lcm, sqrt, log2 and bitwise functions. These functions are implemented for nat, N, BigN, Z, BigZ. See in particular file NPeano for new functions about nat.
The definition of types positive, N, Z is now in file BinNums.v
Major reorganization of ZArith. The initial file ZArith/BinInt.v now contains an internal module Z implementing the Numbers interface for integers. This module Z regroups:
all functions over type Z : Z.add, Z.mul, ...
the minimal proofs of specifications for these functions : Z.add_0_l, ...
an instantiation of all derived properties proved generically in Numbers : Z.add_comm, Z.add_assoc, ...
A large part of ZArith is now simply compatibility notations, for instance Zplus_comm is an alias for Z.add_comm. The direct use of module Z is now recommended instead of relying on these compatibility notations.
Similar major reorganization of NArith, via a module N in NArith/BinNat.v
Concerning the positive datatype, BinPos.v is now in a specific directory PArith, and contains an internal submodule Pos. We regroup there functions such as Pos.add Pos.mul etc as well as many results about them. These results are here proved directly (no Number interface for strictly positive numbers).
Note that in spite of the compatibility layers, all these reorganizations may induce some marginal incompatibilies in scripts. In particular:
the "?=" notation for positive now refers to a binary function Pos.compare, instead of the infamous ternary Pcompare (now Pos.compare_cont).
some hypothesis names generated by the system may changed (typically for a "destruct Z_le_gt_dec") since naming is done after the short name of the head predicate (here now "le" in module Z instead of "Zle", etc).
the internals of Z.add has changed, now relying of Z.pos_sub.
Also note these new notations:
"<?" "<=?" "=?" for boolean tests such as Z.ltb Z.leb Z.eqb.
"÷" for the alternative integer division Z.quot implementing the Truncate convention (former ZOdiv), while the notation for the Coq usual division Z.div implementing the Flooring convention remains "/". Their corresponding modulo functions are Z.rem (no notations) for Z.quot and Z.modulo (infix "mod" notation) for Z.div.
Lemmas about conversions between these datatypes are also organized in modules, see for instance modules Z2Nat, N2Z, etc.
When creating BigN, the macro-generated part NMake_gen is much smaller. The generic part NMake has been reworked and improved. Some changes may introduce incompatibilities. In particular, the order of the arguments for BigN.shiftl and BigN.shiftr is now reversed: the number to shift now comes first. By default, the power function now takes two BigN.
Creation of Vector, an independent library for lists indexed by their length. Vectors' names override lists' one so you should not "Import" the library. All old names changed: function names follow the ocaml ones and, for example, Vcons becomes Vector.cons. You can get [..;..;..]-style notations by importing Vector.VectorNotations.
Removal of TheoryList. Requiring List instead should work most of the time.
New syntax "rew Heq in H" and "rew <- Heq in H" for eq_rect and eq_rect_r (available by importing module EqNotations).
Wf.iter_nat is now Peano.nat_iter (with an implicit type argument).
Internal infrastructure
Opaque proofs are now loaded lazily by default. This allows to be almost as fast as -dont-load-proofs, while being safer (no creation of axioms) and avoiding feature restrictions (Print and Print Assumptions work ok).
Revised hash-consing code allowing more sharing of memory
Experimental support added for camlp4 (the one provided alongside ocaml), simply pass option -usecamlp4 to ./configure. By default camlp5 is used.
Revised build system: no more stages in Makefile thanks to some recursive aspect of recent gnu make, use of vo.itarget files containing .v to compile for both make and ocamlbuild, etc.
Support of cross-compilation via mingw from unix toward Windows, contact P. Letouzey for more informations.
New Makefile rules mli-doc to make html of mli in dev/doc/html and full-stdlib to get a (huge) pdf reflecting the whole standard library.
Extraction
By default, opaque terms are now truly considered opaque by extraction: instead of accessing their body, they are now considered as axioms. The previous behavior can be reactivated via the option "Set Extraction AccessOpaque".
The pretty-printer for Haskell now produces layout-independent code
A new command "Separate Extraction cst1 cst2 ..." that mixes a minimal extracted environment a la "Recursive Extraction" and the production of several files (one per coq source) a la "Extraction Library" (DOC TODO).
New option "Set/Unset Extraction KeepSingleton" for preventing the extraction to optimize singleton container types (DOC TODO).
The extraction now identifies and properly rejects a particular case of universe polymorphism it cannot handle yet (the pair (I,I) being Prop).
Support of anonymous fields in record (#2555).
CoqIDE
CoqIDE now runs coqtop as separated process, making it more robust: coqtop subprocess can be interrupted, or even killed and relaunched (cf button "Restart Coq", ex-"Go to Start"). For allowing such interrupts, the Windows version of coqide now requires Windows >= XP SP1.
The communication between CoqIDE and coqtop is now done via a dialect of XML (DOC TODO).
The backtrack engine of CoqIDE has been reworked, it now uses the "Backtrack" command similarly to Proof General.
The CoqIDE parsing of sentences has be reworked and now supports tactic delimitation via { }.
CoqIDE now accepts the Abort command (wish #2357).
CoqIDE can read coq_makefile files as "project file" and use it to set automatically options to send to coqtop.
Preference files have moved to $XDG_CONFIG_HOME/coq and accelerators are not stored as a list anymore.
Tools
Coq now searches directories specified in COQPATH, $XDG_DATA_HOME/coq, $XDG_DATA_DIRS/coq, and user-contribs before the standard library.
Coq rc file has moved to $XDG_CONFIG_HOME/coq.
Major changes to coq_makefile:
mli/mlpack/mllib taken into account, ml not preproccessed anymore, ml4 work;
mlihtml generates doc of mli, install-doc install the html doc in DOCDIR with the same policy as vo in COQLIB;
More variables are given by coqtop -config, others are defined only if the users doesn't have defined them elsewhere. Consequently, generated makefile should work directly on any architecture;
Packagers can take advantage of $(DSTROOT) introduction. Installation can be made in $XDG_DATA_HOME/coq;
-arg option allows to send option as argument to coqc.
Details of changes in 8.4beta2¶
Commands
Commands "Back" and "BackTo" are now handling the proof states. They may perform some extra steps of backtrack to avoid states where the proof state is unavailable (typically a closed proof).
The commands "Suspend" and "Resume" have been removed.
A basic Show Script has been reintroduced (no indentation).
New command "Set Parsing Explicit" for deactivating parsing (and printing) of implicit arguments (useful for teaching).
New command "Grab Existential Variables" to transform the unresolved evars at the end of a proof into goals.
Tactics
Still no general "info" tactical, but new specific tactics info_auto, info_eauto, info_trivial which provides information on the proofs found by auto/eauto/trivial. Display of these details could also be activated by "Set Info Auto"/"Set Info Eauto"/"Set Info Trivial".
Details on everything tried by auto/eauto/trivial during a proof search could be obtained by "debug auto", "debug eauto", "debug trivial" or by a global "Set Debug Auto"/"Set Debug Eauto"/"Set Debug Trivial".
New command "r string" in Ltac debugger that interprets "idtac string" in Ltac code as a breakpoint and jumps to its next use.
Tactics from the Dp plugin (simplify, ergo, yices, cvc3, z3, cvcl, harvey, zenon, gwhy) have been removed, since Why2 has not been maintained for the last few years. The Why3 plugin should be a suitable replacement in most cases.
Libraries
MSetRBT: a new implementation of MSets via Red-Black trees (initial contribution by Andrew Appel).
MSetAVL: for maximal sharing with the new MSetRBT, the argument order of Node has changed (this should be transparent to regular MSets users).
Module System
The names of modules (and module types) are now in a fully separated namespace from ordinary definitions: "Definition E:=0. Module E. End E." is now accepted.
CoqIDE
CoqIDE now supports the "Restart" command, and "Undo" (with a warning). Better support for "Abort".
Details of changes in 8.4¶
Commands
The "Reset" command is now supported again in files given to coqc or Load.
"Show Script" now indents again the displayed scripts. It can also work correctly across Load'ed files if the option "Unset Atomic Load" is used.
"Open Scope" can now be given the delimiter (e.g. Z) instead of the full scope name (e.g. Z_scope).
Notations
Most compatibility notations of the standard library are now tagged as (compat xyz), where xyz is a former Coq version, for instance "8.3". These notations behave as (only parsing) notations, except that they may triggers warnings (or errors) when used while Coq is not in a corresponding -compat mode.
To activate these compatibility warnings, use "Set Verbose Compat Notations" or the command-line flag -verbose-compat-notations.
For a strict mode without these compatibility notations, use "Unset Compat Notations" or the command-line flag -no-compat-notations.
Tactics
An annotation "eqn:H" or "eqn:?" can be added to a "destruct" or "induction" to make it generate equations in the spirit of "case_eq". The former syntax "_eqn" is discontinued.
The name of the hypothesis introduced by tactic "remember" can be set via the new syntax "remember t as x eqn:H" (wish #2489).
Libraries
Reals: changed definition of PI, no more axiom about sin(PI/2).
SetoidPermutation: a notion of permutation for lists modulo a setoid equality.
BigN: fixed the ocaml code doing the parsing/printing of big numbers.
List: a couple of lemmas added especially about no-duplication, partitions.
Init: Removal of the coercions between variants of sigma-types and subset types (possible source of incompatibility).
Version 8.3¶
Summary of changes¶
Coq version 8.3 is before all a transition version with refinements or extensions of the existing features and libraries and a new tactic nsatz based on Hilbert’s Nullstellensatz for deciding systems of equations over rings.
With respect to libraries, the main evolutions are due to Pierre Letouzey with a rewriting of the library of finite sets FSets and a new round of evolutions in the modular development of arithmetic (library Numbers). The reason for making FSets evolve is that the computational and logical contents were quite intertwined in the original implementation, leading in some cases to longer computations than expected and this problem is solved in the new MSets implementation. As for the modular arithmetic library, it was only dealing with the basic arithmetic operators in the former version and its current extension adds the standard theory of the division, min and max functions, all made available for free to any implementation of \(\mathbb{N}\), \(\mathbb{Z}\) or \(\mathbb{Z}/n\mathbb{Z}\).
The main other evolutions of the library are due to Hugo Herbelin who made a revision of the sorting library (including a certified merge-sort) and to Guillaume Melquiond who slightly revised and cleaned up the library of reals.
The module system evolved significantly. Besides the resolution of some
efficiency issues and a more flexible construction of module types, Élie
Soubiran brought a new model of name equivalence, the
\(\Delta\)-equivalence, which respects as much as possible the names
given by the users. He also designed with Pierre Letouzey a new,
convenient operator <+
for nesting functor application that
provides a light notation for inheriting the properties of cascading
modules.
The new tactic nsatz is due to Loïc Pottier. It works by computing Gröbner bases. Regarding the existing tactics, various improvements have been done by Matthieu Sozeau, Hugo Herbelin and Pierre Letouzey.
Matthieu Sozeau extended and refined the typeclasses and Program features (the Russell language). Pierre Letouzey maintained and improved the extraction mechanism. Bruno Barras and Élie Soubiran maintained the Coq checker, Julien Forest maintained the Function mechanism for reasoning over recursively defined functions. Matthieu Sozeau, Hugo Herbelin and Jean-Marc Notin maintained coqdoc. Frédéric Besson maintained the Micromega platform for deciding systems of inequalities. Pierre Courtieu maintained the support for the Proof General Emacs interface. Claude Marché maintained the plugin for calling external provers (dp). Yves Bertot made some improvements to the libraries of lists and integers. Matthias Puech improved the search functions. Guillaume Melquiond usefully contributed here and there. Yann Régis-Gianas grounded the support for Unicode on a more standard and more robust basis.
Though invisible from outside, Arnaud Spiwack improved the general process of management of existential variables. Pierre Letouzey and Stéphane Glondu improved the compilation scheme of the Coq archive. Vincent Gross provided support to CoqIDE. Jean-Marc Notin provided support for benchmarking and archiving.
Many users helped by reporting problems, providing patches, suggesting improvements or making useful comments, either on the bug tracker or on the Coq-Club mailing list. This includes but not exhaustively Cédric Auger, Arthur Charguéraud, François Garillot, Georges Gonthier, Robin Green, Stéphane Lescuyer, Eelis van der Weegen, ...
Though not directly related to the implementation, special thanks are going to Yves Bertot, Pierre Castéran, Adam Chlipala, and Benjamin Pierce for the excellent teaching materials they provided.
Details of changes¶
Rewriting tactics
Tactic "rewrite" now supports rewriting on ad hoc equalities such as eq_true.
"Hint Rewrite" now checks that the lemma looks like an equation.
New tactic "etransitivity".
Support for heterogeneous equality (JMeq) in "injection" and "discriminate".
Tactic "subst" now supports heterogeneous equality and equality proofs that are dependent (use "simple subst" for preserving compatibility).
Added support for Leibniz-rewriting of dependent hypotheses.
Renamed "Morphism" into "Proper" and "respect" into "proper_prf" (possible source of incompatibility). A partial fix is to define "Notation Morphism R f := (Proper (R%signature) f)."
New tactic variants "rewrite* by" and "autorewrite*" that rewrite respectively the first and all matches whose side-conditions are solved.
"Require Import Setoid" does not export all of "Morphisms" and "RelationClasses" anymore (possible source of incompatibility, fixed by importing "Morphisms" too).
Support added for using Chung-Kil Hur's Heq library for rewriting over heterogeneous equality (courtesy of the library's author).
Tactic "replace" supports matching terms with holes.
Automation tactics
Tactic
intuition
now preserves inneriff
andnot
(exceptional source of incompatibilities solvable by redefiningintuition
asunfold iff, not in *; intuition
, or, for iff only, by usingSet Intuition Iff Unfolding
.)Tactic
tauto
now proves classical tautologies as soon as classical logic (i.e. libraryClassical_Prop
orClassical
) is loaded.Tactic
gappa
has been removed from the Dp plugin.Tactic
firstorder
now supports the combination of itsusing
andwith
options.New
Hint Resolve ->
(or<-
) for declaring iff's as oriented hints (wish #2104).An inductive type as argument of the
using
option ofauto
/eauto
/firstorder
is interpreted as using the collection of its constructors.New decision tactic "nsatz" to prove polynomial equations by computation of Groebner bases.
Other tactics
Tactic "discriminate" now performs intros before trying to discriminate an hypothesis of the goal (previously it applied intro only if the goal had the form t1<>t2) (exceptional source of incompatibilities - former behavior can be obtained by "Unset Discriminate Introduction").
Tactic "quote" now supports quotation of arbitrary terms (not just the goal).
Tactic "idtac" now displays its "list" arguments.
New introduction patterns "*" for introducing the next block of dependent variables and "**" for introducing all quantified variables and hypotheses.
Pattern Unification for existential variables activated in tactics and new option "Unset Tactic Evars Pattern Unification" to deactivate it.
Resolution of canonical structure is now part of the tactic's unification algorithm.
New tactic "decide lemma with hyp" for rewriting decidability lemmas when one knows which side is true.
Improved support of dependent goals over objects in dependent types for "destruct" (rare source of incompatibility that can be avoided by unsetting option "Dependent Propositions Elimination").
Tactic "exists", "eexists", "destruct" and "edestruct" supports iteration using comma-separated arguments.
Tactic names "case" and "elim" now support clauses "as" and "in" and become then synonymous of "destruct" and "induction" respectively.
A new tactic name "exfalso" for the use of 'ex-falso quodlibet' principle. This tactic is simply a shortcut for "elimtype False".
Made quantified hypotheses get the name they would have if introduced in the context (possible but rare source of incompatibilities).
When applying a component of a conjunctive lemma, "apply in" (and sequences of "apply in") now leave the side conditions of the lemmas uniformly after the main goal (possible source of rare incompatibilities).
In "simpl c" and "change c with d", c can be a pattern.
Tactic "revert" now preserves let-in's making it the exact inverse of "intro".
New tactics "clear dependent H" and "revert dependent H" that clears (resp. reverts) H and all the hypotheses that depend on H.
Ltac's pattern-matching now supports matching metavariables that depend on variables bound upwards in the pattern.
Tactic definitions
Ltac definitions support Local option for non-export outside modules.
Support for parsing non-empty lists with separators in tactic notations.
New command "Locate Ltac" to get the full name of an Ltac definition.
Notations
Record syntax
{|x=...; y=...|}
now works inside patterns too.Abbreviations from non-imported module now invisible at printing time.
Abbreviations now use implicit arguments and arguments scopes for printing.
Abbreviations to pure names now strictly behave like the name they refer to (make redirections of qualified names easier).
Abbreviations for applied constant now propagate the implicit arguments and arguments scope of the underlying reference (possible source of incompatibilities generally solvable by changing such abbreviations from e.g.
Notation foo' := (foo x)
toNotation foo' y := (foo x (y:=y))
).The "where" clause now supports multiple notations per defined object.
Recursive notations automatically expand one step on the left for better factorization; recursion notations inner separators now ensured being tokens.
Added "Reserved Infix" as a specific shortcut of the corresponding "Reserved Notation".
Open/Close Scope command supports Global option in sections.
Specification language
New support for local binders in the syntax of Record/Structure fields.
Fixpoint/CoFixpoint now support building part or all of bodies using tactics.
Binders given before ":" in lemmas and in definitions built by tactics are now automatically introduced (possible source of incompatibility that can be resolved by invoking "Unset Automatic Introduction").
New support for multiple implicit arguments signatures per reference.
Module system
Include Type is now deprecated since Include now accepts both modules and module types.
Declare ML Module supports Local option.
The sharing between non-logical object and the management of the name-space has been improved by the new "Delta-equivalence" on qualified name.
The include operator has been extended to high-order structures
Sequences of Include can be abbreviated via new syntax "<+".
A module (or module type) can be given several "<:" signatures.
Interactive proofs are now permitted in module type. Functors can hence be declared as Module Type and be used later to type themselves.
A functor application can be prefixed by a "!" to make it ignore any "Inline" annotation in the type of its argument(s) (for examples of use of the new features, see libraries Structures and Numbers).
Coercions are now active only when modules are imported (use "Set Automatic Coercions Import" to get the behavior of the previous versions of Coq).
Extraction
When using (Recursive) Extraction Library, the filenames are directly the Coq ones with new appropriate extensions : we do not force anymore uncapital first letters for Ocaml and capital ones for Haskell.
The extraction now tries harder to avoid code transformations that can be dangerous for the complexity. In particular many eta-expansions at the top of functions body are now avoided, clever partial applications will likely be preserved, let-ins are almost always kept, etc.
In the same spirit, auto-inlining is now disabled by default, except for induction principles, since this feature was producing more frequently weird code than clear gain. The previous behavior can be restored via "Set Extraction AutoInline".
Unicode characters in identifiers are now transformed into ascii strings that are legal in Ocaml and other languages.
Harsh support of module extraction to Haskell and Scheme: module hierarchy is flattened, module abbreviations and functor applications are expanded, module types and unapplied functors are discarded.
Less unsupported situations when extracting modules to Ocaml. In particular module parameters might be alpha-renamed if a name clash is detected.
Extract Inductive is now possible toward non-inductive types (e.g. nat => int)
Extraction Implicit: this new experimental command allows to mark some arguments of a function or constructor for removed during extraction, even if these arguments don't fit the usual elimination principles of extraction, for instance the length n of a vector.
Files ExtrOcaml*.v in plugins/extraction try to provide a library of common extraction commands: mapping of basics types toward Ocaml's counterparts, conversions from/to int and big_int, or even complete mapping of nat,Z,N to int or big_int, or mapping of ascii to char and string to char list (in this case recognition of ascii constants is hard-wired in the extraction).
Program
Streamlined definitions using well-founded recursion and measures so that they can work on any subset of the arguments directly (uses currying).
Try to automatically clear structural fixpoint prototypes in obligations to avoid issues with opacity.
Use return type clause inference in pattern-matching as in the standard typing algorithm.
Support [Local Obligation Tactic] and [Next Obligation with tactic].
Use [Show Obligation Tactic] to print the current default tactic.
[fst] and [snd] have maximal implicit arguments in Program now (possible source of incompatibility).
Type classes
Declaring axiomatic type class instances in Module Type should be now done via new command "Declare Instance", while the syntax "Instance" now always provides a concrete instance, both in and out of Module Type.
Use [Existing Class foo] to declare a preexisting object [foo] as a class. [foo] can be an inductive type or a constant definition. No projections or instances are defined.
Various bug fixes and improvements: support for defined fields, anonymous instances, declarations giving terms, better handling of sections and [Context].
Commands
New command "Timeout <n> <command>." interprets a command and a timeout interrupts the execution after <n> seconds.
New command "Compute <expr>." is a shortcut for "Eval vm_compute in <expr>".
New command "Fail <command>." interprets a command and is successful iff the command fails on an error (but not an anomaly). Handy for tests and illustration of wrong commands.
Most commands referring to constant (e.g. Print or About) now support referring to the constant by a notation string.
New option "Boolean Equality Schemes" to make generation of boolean equality automatic for datatypes (together with option "Decidable Equality Schemes", this replaces deprecated option "Equality Scheme").
Made support for automatic generation of case analysis schemes available to user (governed by option "Set Case Analysis Schemes").
New command
Global? Generalizable AllNo VariableVariables ident*
to declare which identifiers are generalizable in ``{} `
and ``() `
binders.New command "Print Opaque Dependencies" to display opaque constants in addition to all variables, parameters or axioms a theorem or definition relies on.
New command "Declare Reduction <id> := <conv_expr>", allowing to write later "Eval <id> in ...". This command accepts a Local variant.
Syntax of Implicit Type now supports more than one block of variables of a given type.
Command "Canonical Structure" now warns when it has no effects.
Commands of the form "Set X" or "Unset X" now support "Local" and "Global" prefixes.
Library
Use "standard" Coq names for the properties of eq and identity (e.g. refl_equal is now eq_refl). Support for compatibility is provided.
The function Compare_dec.nat_compare is now defined directly, instead of relying on lt_eq_lt_dec. The earlier version is still available under the name nat_compare_alt.
Lemmas in library Relations and Reals have been homogenized a bit.
The implicit argument of Logic.eq is now maximally inserted, allowing to simply write "eq" instead of "@eq _" in morphism signatures.
Wrongly named lemmas (Zlt_gt_succ and Zlt_succ_gt) fixed (potential source of incompatibilities)
List library:
Definitions of list, length and app are now in Init/Datatypes. Support for compatibility is provided.
Definition of Permutation is now in Sorting/Permtation.v
Some other light revisions and extensions (possible source of incompatibilities solvable by qualifying names accordingly).
In ListSet, set_map has been fixed (source of incompatibilities if used).
Sorting library:
new mergesort of worst-case complexity O(n*ln(n)) made available in Mergesort.v;
former notion of permutation up to setoid from Permutation.v is deprecated and moved to PermutSetoid.v;
heapsort from Heap.v of worst-case complexity O(n*n) is deprecated;
new file Sorted.v for some definitions of being sorted.
Structure library. This new library is meant to contain generic structures such as types with equalities or orders, either in Module version (for now) or Type Classes (still to do):
DecidableType.v and OrderedType.v: initial notions for FSets/FMaps, left for compatibility but considered as deprecated.
Equalities.v and Orders.v: evolutions of the previous files, with fine-grain Module architecture, many variants, use of Equivalence and other relevant Type Classes notions.
OrdersTac.v: a generic tactic for solving chains of (in)equalities over variables. See {Nat,N,Z,P}OrderedType.v for concrete instances.
GenericMinMax.v: any ordered type can be equipped with min and max. We derived here all the generic properties of these functions.
MSets library: an important evolution of the FSets library. "MSets" stands for Modular (Finite) Sets, by contrast with a forthcoming library of Class (Finite) Sets contributed by S. Lescuyer which will be integrated with the next release of Coq. The main features of MSets are:
The use of Equivalence, Proper and other Type Classes features easing the handling of setoid equalities.
The interfaces are now stated in iff-style. Old specifications are now derived properties.
The compare functions are now pure, and return a "comparison" value. Thanks to the CompSpec inductive type, reasoning on them remains easy.
Sets structures requiring invariants (i.e. sorted lists) are built first as "Raw" sets (pure objects and separate proofs) and attached with their proofs thanks to a generic functor. "Raw" sets have now a proper interface and can be manipulated directly.
Note: No Maps yet in MSets. The FSets library is still provided for compatibility, but will probably be considered as deprecated in the next release of Coq.
Numbers library:
The abstract layer (NatInt, Natural/Abstract, Integer/Abstract) has been simplified and enhance thanks to new features of the module system such as Include (see above). It has been extended to Euclidean division (three flavors for integers: Trunc, Floor and Math).
The arbitrary-large efficient numbers (BigN, BigZ, BigQ) has also been reworked. They benefit from the abstract layer improvements (especially for div and mod). Note that some specifications have slightly changed (compare, div, mod, shift{r,l}). Ring/Field should work better (true recognition of constants).
Tools
Option -R now supports binding Coq root read-only.
New coqtop/coqc option -beautify to reformat .v files (usable e.g. to globally update notations).
New tool beautify-archive to beautify a full archive of developments.
New coqtop/coqc option -compat X.Y to simulate the general behavior of previous versions of Coq (provides e.g. support for 8.2 compatibility).
Coqdoc
List have been revamped. List depth and scope is now determined by an "offside" whitespace rule.
Text may be italicized by placing it in _underscores_.
The "--index <string>" flag changes the filename of the index.
The "--toc-depth <int>" flag limits the depth of headers which are included in the table of contents.
The "--lib-name <string>" flag prints "<string> Foo" instead of "Library Foo" where library titles are called for. The "--no-lib-name" flag eliminates the extra title.
New option "--parse-comments" to allow parsing of regular
(* *)
comments.New option "--plain-comments" to disable interpretation inside comments.
New option "--interpolate" to try and typeset identifiers in Coq escapings using the available globalization information.
New option "--external url root" to refer to external libraries.
Links to section variables and notations now supported.
Internal infrastructure
To avoid confusion with the repository of user's contributions, the subdirectory "contrib" has been renamed into "plugins". On platforms supporting ocaml native dynlink, code located there is built as loadable plugins for coqtop.
An experimental build mechanism via ocamlbuild is provided. From the top of the archive, run ./configure as usual, and then ./build. Feedback about this build mechanism is most welcome. Compiling Coq on platforms such as Windows might be simpler this way, but this remains to be tested.
The Makefile system has been simplified and factorized with the ocamlbuild system. In particular "make" takes advantage of .mllib files for building .cma/.cmxa. The .vo files to compile are now listed in several vo.itarget files.
Version 8.2¶
Summary of changes¶
Coq version 8.2 adds new features, new libraries and improves on many various aspects.
Regarding the language of Coq, the main novelty is the introduction by Matthieu Sozeau of a package of commands providing Haskell-style typeclasses. Typeclasses, which come with a few convenient features such as type-based resolution of implicit arguments, play a new landmark role in the architecture of Coq with respect to automation. For instance, thanks to typeclass support, Matthieu Sozeau could implement a new resolution-based version of the tactics dedicated to rewriting on arbitrary transitive relations.
Another major improvement of Coq 8.2 is the evolution of the arithmetic libraries and of the tools associated with them. Benjamin Grégoire and Laurent Théry contributed a modular library for building arbitrarily large integers from bounded integers while Evgeny Makarov contributed a modular library of abstract natural and integer arithmetic together with a few convenient tactics. On his side, Pierre Letouzey made numerous extensions to the arithmetic libraries on \(\mathbb{Z}\) and \(\mathbb{Q}\), including extra support for automation in presence of various number-theory concepts.
Frédéric Besson contributed a reflective tactic based on Krivine-Stengle Positivstellensatz (the easy way) for validating provability of systems of inequalities. The platform is flexible enough to support the validation of any algorithm able to produce a “certificate” for the Positivstellensatz and this covers the case of Fourier-Motzkin (for linear systems in \(\mathbb{Q}\) and \(\mathbb{R}\)), Fourier-Motzkin with cutting planes (for linear systems in \(\mathbb{Z}\)) and sum-of-squares (for non-linear systems). Evgeny Makarov made the platform generic over arbitrary ordered rings.
Arnaud Spiwack developed a library of 31-bits machine integers and, relying on Benjamin Grégoire and Laurent Théry’s library, delivered a library of unbounded integers in base \(2^{31}\). As importantly, he developed a notion of “retro-knowledge” so as to safely extend the kernel-located bytecode-based efficient evaluation algorithm of Coq version 8.1 to use 31-bits machine arithmetic for efficiently computing with the library of integers he developed.
Beside the libraries, various improvements were contributed to provide a more comfortable end-user language and more expressive tactic language. Hugo Herbelin and Matthieu Sozeau improved the pattern matching compilation algorithm (detection of impossible clauses in pattern matching, automatic inference of the return type). Hugo Herbelin, Pierre Letouzey and Matthieu Sozeau contributed various new convenient syntactic constructs and new tactics or tactic features: more inference of redundant information, better unification, better support for proof or definition by fixpoint, more expressive rewriting tactics, better support for meta-variables, more convenient notations...
Élie Soubiran improved the module system, adding new features (such as an “include” command) and making it more flexible and more general. He and Pierre Letouzey improved the support for modules in the extraction mechanism.
Matthieu Sozeau extended the Russell language, ending in an convenient way to write programs of given specifications, Pierre Corbineau extended the Mathematical Proof Language and the automation tools that accompany it, Pierre Letouzey supervised and extended various parts of the standard library, Stéphane Glondu contributed a few tactics and improvements, Jean-Marc Notin provided help in debugging, general maintenance and coqdoc support, Vincent Siles contributed extensions of the Scheme command and of injection.
Bruno Barras implemented the coqchk
tool: this is a stand-alone
type checker that can be used to certify .vo files. Especially, as this
verifier runs in a separate process, it is granted not to be “hijacked”
by virtually malicious extensions added to Coq.
Yves Bertot, Jean-Christophe Filliâtre, Pierre Courtieu and Julien Forest acted as maintainers of features they implemented in previous versions of Coq.
Julien Narboux contributed to CoqIDE. Nicolas Tabareau made the adaptation of the interface of the old “setoid rewrite” tactic to the new version. Lionel Mamane worked on the interaction between Coq and its external interfaces. With Samuel Mimram, he also helped making Coq compatible with recent software tools. Russell O’Connor, Cezary Kaliszyk, Milad Niqui contributed to improve the libraries of integers, rational, and real numbers. We also thank many users and partners for suggestions and feedback, in particular Pierre Castéran and Arthur Charguéraud, the INRIA Marelle team, Georges Gonthier and the INRIA-Microsoft Mathematical Components team, the Foundations group at Radboud university in Nijmegen, reporters of bugs and participants to the Coq-Club mailing list.
Details of changes¶
Language
If a fixpoint is not written with an explicit { struct ... }, then all arguments are tried successively (from left to right) until one is found that satisfies the structural decreasing condition.
New experimental typeclass system giving ad-hoc polymorphism and overloading based on dependent records and implicit arguments.
New syntax "let 'pat := b in c" for let-binding using irrefutable patterns.
New syntax "forall {A}, T" for specifying maximally inserted implicit arguments in terms.
Sort of Record/Structure, Inductive and CoInductive defaults to Type if omitted.
(Co)Inductive types can be defined as records (e.g. "CoInductive stream := { hd : nat; tl : stream }.")
New syntax "Theorem id1:t1 ... with idn:tn" for proving mutually dependent statements.
Support for sort-polymorphism on constants denoting inductive types.
Several evolutions of the module system (handling of module aliases, functorial module types, an Include feature, etc).
Prop now a subtype of Set (predicative and impredicative forms).
Recursive inductive types in Prop with a single constructor of which all arguments are in Prop is now considered to be a singleton type. It consequently supports all eliminations to Prop, Set and Type. As a consequence, Acc_rect has now a more direct proof [possible source of easily fixed incompatibility in case of manual definition of a recursor in a recursive singleton inductive type].
Commands
Added option Global to "Arguments Scope" for section surviving.
Added option "Unset Elimination Schemes" to deactivate the automatic generation of elimination schemes.
Modification of the Scheme command so you can ask for the name to be automatically computed (e.g. Scheme Induction for nat Sort Set).
New command "Combined Scheme" to build combined mutual induction principles from existing mutual induction principles.
New command "Scheme Equality" to build a decidable (boolean) equality for simple inductive datatypes and a decision property over this equality (e.g. Scheme Equality for nat).
Added option "Set Equality Scheme" to make automatic the declaration of the boolean equality when possible.
Source of universe inconsistencies now printed when option "Set Printing Universes" is activated.
New option "Set Printing Existential Instances" for making the display of existential variable instances explicit.
Support for option "[id1 ... idn]", and "-[id1 ... idn]", for the "compute"/"cbv" reduction strategy, respectively meaning reduce only, or everything but, the constants id1 ... idn. "lazy" alone or followed by "[id1 ... idn]", and "-[id1 ... idn]" also supported, meaning apply all of beta-iota-zeta-delta, possibly restricting delta.
New command "Strategy" to control the expansion of constants during conversion tests. It generalizes commands Opaque and Transparent by introducing a range of levels. Lower levels are assigned to constants that should be expanded first.
New options Global and Local to Opaque and Transparent.
New command "Print Assumptions" to display all variables, parameters or axioms a theorem or definition relies on.
"Add Rec LoadPath" now provides references to libraries using partially qualified names (this holds also for coqtop/coqc option -R).
SearchAbout supports negated search criteria, reference to logical objects by their notation, and more generally search of subterms.
"Declare ML Module" now allows to import .cmxs files when Coq is compiled in native code with a version of OCaml that supports native Dynlink (>= 3.11).
Specific sort constraints on Record now taken into account.
"Print LoadPath" supports a path argument to filter the display.
Libraries
Several parts of the libraries are now in Type, in particular FSets, SetoidList, ListSet, Sorting, Zmisc. This may induce a few incompatibilities. In case of trouble while fixing existing development, it may help to simply declare Set as an alias for Type (see file SetIsType).
New arithmetical library in theories/Numbers. It contains:
an abstract modular development of natural and integer arithmetics in Numbers/Natural/Abstract and Numbers/Integer/Abstract
an implementation of efficient computational bounded and unbounded integers that can be mapped to processor native arithmetics. See Numbers/Cyclic/Int31 for 31-bit integers and Numbers/Natural/BigN for unbounded natural numbers and Numbers/Integer/BigZ for unbounded integers.
some proofs that both older libraries Arith, ZArith and NArith and newer BigN and BigZ implement the abstract modular development. This allows in particular BigN and BigZ to already come with a large database of basic lemmas and some generic tactics (ring),
This library has still an experimental status, as well as the processor-acceleration mechanism, but both its abstract and its concrete parts are already quite usable and could challenge the use of nat, N and Z in actual developments. Moreover, an extension of this framework to rational numbers is ongoing, and an efficient Q structure is already provided (see Numbers/Rational/BigQ), but this part is currently incomplete (no abstract layer and generic lemmas).
Many changes in FSets/FMaps. In practice, compatibility with earlier version should be fairly good, but some adaptations may be required.
Interfaces of unordered ("weak") and ordered sets have been factorized thanks to new features of Coq modules (in particular Include), see FSetInterface. Same for maps. Hints in these interfaces have been reworked (they are now placed in a "set" database).
To allow full subtyping between weak and ordered sets, a field "eq_dec" has been added to OrderedType. The old version of OrderedType is now called MiniOrderedType and functor MOT_to_OT allow to convert to the new version. The interfaces and implementations of sets now contain also such a "eq_dec" field.
FSetDecide, contributed by Aaron Bohannon, contains a decision procedure allowing to solve basic set-related goals (for instance, is a point in a particular set ?). See FSetProperties for examples.
Functors of properties have been improved, especially the ones about maps, that now propose some induction principles. Some properties of fold need less hypothesis.
More uniformity in implementations of sets and maps: they all use implicit arguments, and no longer export unnecessary scopes (see bug #1347)
Internal parts of the implementations based on AVL have evolved a lot. The main files FSetAVL and FMapAVL are now much more lightweight now. In particular, minor changes in some functions has allowed to fully separate the proofs of operational correctness from the proofs of well-balancing: well-balancing is critical for efficiency, but not anymore for proving that these trees implement our interfaces, hence we have moved these proofs into appendix files FSetFullAVL and FMapFullAVL. Moreover, a few functions like union and compare have been modified in order to be structural yet efficient. The appendix files also contains alternative versions of these few functions, much closer to the initial Ocaml code and written via the Function framework.
Library IntMap, subsumed by FSets/FMaps, has been removed from Coq Standard Library and moved into a user contribution Cachan/IntMap
Better computational behavior of some constants (eq_nat_dec and le_lt_dec more efficient, Z_lt_le_dec and Positive_as_OT.compare transparent, ...) (exceptional source of incompatibilities).
Boolean operators moved from module Bool to module Datatypes (may need to rename qualified references in script and force notations || and && to be at levels 50 and 40 respectively).
The constructors xI and xO of type positive now have postfix notations "~1" and "~0", allowing to write numbers in binary form easily, for instance 6 is 1~1~0 and 4*p is p~0~0 (see BinPos.v).
Improvements to NArith (Nminus, Nmin, Nmax), and to QArith (in particular a better power function).
Changes in ZArith: several additional lemmas (used in theories/Numbers), especially in Zdiv, Znumtheory, Zpower. Moreover, many results in Zdiv have been generalized: the divisor may simply be non-null instead of strictly positive (see lemmas with name ending by "_full"). An alternative file ZOdiv proposes a different behavior (the one of Ocaml) when dividing by negative numbers.
Changes in Arith: EqNat and Wf_nat now exported from Arith, some constructions on nat that were outside Arith are now in (e.g. iter_nat).
In SetoidList, eqlistA now expresses that two lists have similar elements at the same position, while the predicate previously called eqlistA is now equivlistA (this one only states that the lists contain the same elements, nothing more).
Changes in Reals:
Most statement in "sigT" (including the completeness axiom) are now in "sig" (in case of incompatibility, use proj1_sig instead of projT1, sig instead of sigT, etc).
More uniform naming scheme (identifiers in French moved to English, consistent use of 0 -- zero -- instead of O -- letter O --, etc).
Lemma on prod_f_SO is now on prod_f_R0.
Useless hypothesis of ln_exists1 dropped.
New Rlogic.v states a few logical properties about R axioms.
RIneq.v extended and made cleaner.
Slight restructuration of the Logic library regarding choice and classical logic. Addition of files providing intuitionistic axiomatizations of descriptions: Epsilon.v, Description.v and IndefiniteDescription.v.
Definition of pred and minus made compatible with the structural decreasing criterion for use in fixpoints.
Files Relations/Rstar.v and Relations/Newman.v moved out to the user contribution repository (contribution CoC_History). New lemmas about transitive closure added and some bound variables renamed (exceptional risk of incompatibilities).
Syntax for binders in terms (e.g. for "exists") supports anonymous names.
Notations, coercions, implicit arguments and type inference
More automation in the inference of the return clause of dependent pattern-matching problems.
Experimental allowance for omission of the clauses easily detectable as impossible in pattern-matching problems.
Improved inference of implicit arguments.
New options "Set Maximal Implicit Insertion", "Set Reversible Pattern Implicit", "Set Strongly Strict Implicit" and "Set Printing Implicit Defensive" for controlling inference and use of implicit arguments.
New modifier in "Implicit Arguments" to force an implicit argument to be maximally inserted.
New modifier of "Implicit Arguments" to enrich the set of implicit arguments.
New options Global and Local to "Implicit Arguments" for section surviving or non-export outside module.
Level "constr" moved from 9 to 8.
Structure/Record now printed as Record (unless option Printing All is set).
Support for parametric notations defining constants.
Insertion of coercions below product types refrains to unfold constants (possible source of incompatibility).
New support for fix/cofix in notations.
Tactic Language
Second-order pattern-matching now working in Ltac "match" clauses (syntax for second-order unification variable is "@?X").
Support for matching on let bindings in match context using syntax "H := body" or "H := body : type".
Ltac accepts integer arguments (syntax is "ltac:nnn" for nnn an integer).
The general sequence tactical "expr_0 ; [ expr_1 | ... | expr_n ]" is extended so that at most one expr_i may have the form "expr .." or just "..". Also, n can be different from the number of subgoals generated by expr_0. In this case, the value of expr (or idtac in case of just "..") is applied to the intermediate subgoals to make the number of tactics equal to the number of subgoals.
A name used as the name of the parameter of a lemma (like f in "apply f_equal with (f:=t)") is now interpreted as a ltac variable if such a variable exists (this is a possible source of incompatibility and it can be fixed by renaming the variables of a ltac function into names that do not clash with the lemmas parameter names used in the tactic).
New syntax "Ltac tac ::= ..." to rebind a tactic to a new expression.
"let rec ... in ... " now supported for expressions without explicit parameters; interpretation is lazy to the contrary of "let ... in ..."; hence, the "rec" keyword can be used to turn the argument of a "let ... in ..." into a lazy one.
Patterns for hypotheses types in "match goal" are now interpreted in type_scope.
A bound variable whose name is not used elsewhere now serves as metavariable in "match" and it gets instantiated by an identifier (allow e.g. to extract the name of a statement like "exists x, P x").
New printing of Ltac call trace for better debugging.
Tactics
New tactics "apply -> term", "apply <- term", "apply -> term in ident", "apply <- term in ident" for applying equivalences (iff).
Slight improvement of the hnf and simpl tactics when applied on expressions with explicit occurrences of match or fix.
New tactics "eapply in", "erewrite", "erewrite in".
New tactics "ediscriminate", "einjection", "esimplify_eq".
Tactics "discriminate", "injection", "simplify_eq" now support any term as argument. Clause "with" is also supported.
Unfoldable references can be given by notation's string rather than by name in unfold.
The "with" arguments are now typed using informations from the current goal: allows support for coercions and more inference of implicit arguments.
Application of "f_equal"-style lemmas works better.
Tactics elim, case, destruct and induction now support variants eelim, ecase, edestruct and einduction.
Tactics destruct and induction now support the "with" option and the "in" clause option. If the option "in" is used, an equality is added to remember the term to which the induction or case analysis applied (possible source of parsing incompatibilities when destruct or induction is part of a let-in expression in Ltac; extra parentheses are then required).
New support for "as" clause in tactics "apply in" and "eapply in".
Some new intro patterns:
intro pattern "?A" genererates a fresh name based on A. Caveat about a slight loss of compatibility: Some intro patterns don't need space between them. In particular intros ?a?b used to be legal and equivalent to intros ? a ? b. Now it is still legal but equivalent to intros ?a ?b.
intro pattern "(A & ... & Y & Z)" synonym to "(A,....,(Y,Z)))))" for right-associative constructs like /or exists.
Several syntax extensions concerning "rewrite":
"rewrite A,B,C" can be used to rewrite A, then B, then C. These rewrites occur only on the first subgoal: in particular, side-conditions of the "rewrite A" are not concerned by the "rewrite B,C".
"rewrite A by tac" allows to apply tac on all side-conditions generated by the "rewrite A".
"rewrite A at n" allows to select occurrences to rewrite: rewrite only happen at the n-th exact occurrence of the first successful matching of A in the goal.
"rewrite 3 A" or "rewrite 3!A" is equivalent to "rewrite A,A,A".
"rewrite !A" means rewriting A as long as possible (and at least once).
"rewrite 3?A" means rewriting A at most three times.
"rewrite ?A" means rewriting A as long as possible (possibly never).
many of the above extensions can be combined with each other.
Introduction patterns better respect the structure of context in presence of missing or extra names in nested disjunction-conjunction patterns [possible source of rare incompatibilities].
New syntax "rename a into b, c into d" for "rename a into b; rename c into d"
New tactics "dependent induction/destruction H [ generalizing id_1 .. id_n ]" to do induction-inversion on instantiated inductive families à la BasicElim.
Tactics "apply" and "apply in" now able to reason modulo unfolding of constants (possible source of incompatibility in situations where apply may fail, e.g. as argument of a try or a repeat and in a ltac function); versions that do not unfold are renamed into "simple apply" and "simple apply in" (usable for compatibility or for automation).
Tactics "apply" and "apply in" now able to traverse conjunctions and to select the first matching lemma among the components of the conjunction; tactic "apply" also able to apply lemmas of conclusion an empty type.
Tactic "apply" now supports application of several lemmas in a row.
Tactics "set" and "pose" can set functions using notation "(f x1..xn := c)".
New tactic "instantiate" (without argument).
Tactic firstorder "with" and "using" options have their meaning swapped for consistency with auto/eauto (source of incompatibility).
Tactic "generalize" now supports "at" options to specify occurrences and "as" options to name the quantified hypotheses.
New tactic "specialize H with a" or "specialize (H a)" allows to transform in-place a universally-quantified hypothesis (H : forall x, T x) into its instantiated form (H : T a). Nota: "specialize" was in fact there in earlier versions of Coq, but was undocumented, and had a slightly different behavior.
New tactic "contradict H" can be used to solve any kind of goal as long as the user can provide afterwards a proof of the negation of the hypothesis H. If H is already a negation, say ~T, then a proof of T is asked. If the current goal is a negation, say ~U, then U is saved in H afterwards, hence this new tactic "contradict" extends earlier tactic "swap", which is now obsolete.
Tactics f_equal is now done in ML instead of Ltac: it now works on any equality of functions, regardless of the arity of the function.
New options "before id", "at top", "at bottom" for tactics "move"/"intro".
Some more debug of reflexive omega (
romega
), and internal clarifications. Moreover, romega now has a variantromega with *
that can be also used on non-Z goals (nat, N, positive) via a call to a translation tactic named zify (its purpose is to Z-ify your goal...). This zify may also be used independently of romega.Tactic "remember" now supports an "in" clause to remember only selected occurrences of a term.
Tactic "pose proof" supports name overriding in case of specialization of an hypothesis.
Semi-decision tactic "jp" for first-order intuitionistic logic moved to user contributions (subsumed by "firstorder").
Program
Moved useful tactics in theories/Program and documented them.
Add Program.Basics which contains standard definitions for functional programming (id, apply, flip...)
More robust obligation handling, dependent pattern-matching and well-founded definitions.
New syntax " dest term as pat in term " for destructing objects using an irrefutable pattern while keeping equalities (use this instead of "let" in Programs).
Program CoFixpoint is accepted, Program Fixpoint uses the new way to infer which argument decreases structurally.
Program Lemma, Axiom etc... now permit to have obligations in the statement iff they can be automatically solved by the default tactic.
Renamed "Obligations Tactic" command to "Obligation Tactic".
New command "Preterm [ of id ]" to see the actual term fed to Coq for debugging purposes.
New option "Transparent Obligations" to control the declaration of obligations as transparent or opaque. All obligations are now transparent by default, otherwise the system declares them opaque if possible.
Changed the notations "left" and "right" to "in_left" and "in_right" to hide the proofs in standard disjunctions, to avoid breaking existing scripts when importing Program. Also, put them in program_scope.
Type Classes
New "Class", "Instance" and "Program Instance" commands to define classes and instances documented in the reference manual.
New binding construct " [ Class_1 param_1 .. param_n, Class_2 ... ] " for binding type classes, usable everywhere.
New command " Print Classes " and " Print Instances some_class " to print tables for typeclasses.
New default eauto hint database "typeclass_instances" used by the default typeclass instance search tactic.
New theories directory "theories/Classes" for standard typeclasses declarations. Module Classes.RelationClasses is a typeclass port of Relation_Definitions plus a generic development of algebra on n-ary heterogeneous predicates.
Setoid rewriting
Complete (and still experimental) rewrite of the tactic based on typeclasses. The old interface and semantics are almost entirely respected, except:
Import Setoid is now mandatory to be able to call setoid_replace and declare morphisms.
"-->", "++>" and "==>" are now right associative notations declared at level 55 in scope signature_scope. Their introduction may break existing scripts that defined them as notations with different levels.
One needs to use [Typeclasses unfold [cst]] if [cst] is used as an abbreviation hiding products in types of morphisms, e.g. if ones redefines [relation] and declares morphisms whose type mentions [relation].
The [setoid_rewrite]'s semantics change when rewriting with a lemma: it can rewrite two different instantiations of the lemma at once. Use [setoid_rewrite H at 1] for (almost) the usual semantics. [setoid_rewrite] will also try to rewrite under binders now, and can succeed on different terms than before. In particular, it will unify under let-bound variables. When called through [rewrite], the semantics are unchanged though.
[Add Morphism term : id] has different semantics when used with parametric morphism: it will try to find a relation on the parameters too. The behavior has also changed with respect to default relations: the most recently declared Setoid/Relation will be used, the documentation explains how to customize this behavior.
Parametric Relation and Morphism are declared differently, using the new [Add Parametric] commands, documented in the manual.
Setoid_Theory is now an alias to Equivalence, scripts building objects of type Setoid_Theory need to unfold (or "red") the definitions of Reflexive, Symmetric and Transitive in order to get the same goals as before. Scripts which introduced variables explicitly will not break.
The order of subgoals when doing [setoid_rewrite] with side-conditions is always the same: first the new goal, then the conditions.
New standard library modules
Classes.Morphisms
declares standard morphisms onrefl
/sym
/trans
relations.Classes.Morphisms_Prop
declares morphisms on propositional connectives andClasses.Morphisms_Relations
on generalized predicate connectives.Classes.Equivalence
declares notations and tactics related to equivalences andClasses.SetoidTactics
defines the setoid_replace tactics and some support for theAdd *
interface, notably the tactic applied automatically before eachAdd Morphism
proof.User-defined subrelations are supported, as well as higher-order morphisms and rewriting under binders. The tactic is also extensible entirely in Ltac. The documentation has been updated to cover these features.
[setoid_rewrite] and [rewrite] now support the [at] modifier to select occurrences to rewrite, and both use the [setoid_rewrite] code, even when rewriting with leibniz equality if occurrences are specified.
Extraction
Improved behavior of the Caml extraction of modules: name clashes should not happen anymore.
The command Extract Inductive has now a syntax for infix notations. This allows in particular to map Coq lists and pairs onto OCaml ones:
Extract Inductive list => list [ "[]" "(::)" ].
Extract Inductive prod => "(*)" [ "(,)" ].
In pattern matchings, a default pattern "| _ -> ..." is now used whenever possible if several branches are identical. For instance, functions corresponding to decidability of equalities are now linear instead of quadratic.
A new instruction Extraction Blacklist id1 .. idn allows to prevent filename conflits with existing code, for instance when extracting module List to Ocaml.
CoqIDE
CoqIDE font defaults to monospace so as indentation to be meaningful.
CoqIDE supports nested goals and any other kind of declaration in the middle of a proof.
Undoing non-tactic commands in CoqIDE works faster.
New CoqIDE menu for activating display of various implicit informations.
Added the possibility to choose the location of tabs in coqide: (in Edit->Preferences->Misc)
New Open and Save As dialogs in CoqIDE which filter
*.v
files.
Tools
New stand-alone .vo files verifier "coqchk".
Extended -I coqtop/coqc option to specify a logical dir: "-I dir -as coqdir".
New coqtop/coqc option -exclude-dir to exclude subdirs for option -R.
The binary "parser" has been renamed to "coq-parser".
Improved coqdoc and dump of globalization information to give more meta-information on identifiers. All categories of Coq definitions are supported, which makes typesetting trivial in the generated documentation. Support for hyperlinking and indexing developments in the tex output has been implemented as well.
Miscellaneous
Coq installation provides enough files so that Ocaml's extensions need not the Coq sources to be compiled (this assumes O'Caml 3.10 and Camlp5).
New commands "Set Whelp Server" and "Set Whelp Getter" to customize the Whelp search tool.
Syntax of "Test Printing Let ref" and "Test Printing If ref" changed into "Test Printing Let for ref" and "Test Printing If for ref".
An overhauled build system (new Makefiles); see dev/doc/build-system.txt.
Add -browser option to configure script.
Build a shared library for the C part of Coq, and use it by default on non-(Windows or MacOS) systems. Bytecode executables are now pure. The behavior is configurable with -coqrunbyteflags, -coqtoolsbyteflags and -custom configure options.
Complexity tests can be skipped by setting the environment variable COQTEST_SKIPCOMPLEXITY.
Version 8.1¶
Summary of changes¶
Coq version 8.1 adds various new functionalities.
Benjamin Grégoire implemented an alternative algorithm to check the convertibility of terms in the Coq type checker. This alternative algorithm works by compilation to an efficient bytecode that is interpreted in an abstract machine similar to Xavier Leroy’s ZINC machine. Convertibility is performed by comparing the normal forms. This alternative algorithm is specifically interesting for proofs by reflection. More generally, it is convenient in case of intensive computations.
Christine Paulin implemented an extension of inductive types allowing recursively non-uniform parameters. Hugo Herbelin implemented sort-polymorphism for inductive types (now called template polymorphism).
Claudio Sacerdoti Coen improved the tactics for rewriting on arbitrary compatible equivalence relations. He also generalized rewriting to arbitrary transition systems.
Claudio Sacerdoti Coen added new features to the module system.
Benjamin Grégoire, Assia Mahboubi and Bruno Barras developed a new, more efficient and more general simplification algorithm for rings and semirings.
Laurent Théry and Bruno Barras developed a new, significantly more efficient simplification algorithm for fields.
Hugo Herbelin, Pierre Letouzey, Julien Forest, Julien Narboux and Claudio Sacerdoti Coen added new tactic features.
Hugo Herbelin implemented matching on disjunctive patterns.
New mechanisms made easier the communication between Coq and external provers. Nicolas Ayache and Jean-Christophe Filliâtre implemented connections with the provers cvcl, Simplify and zenon. Hugo Herbelin implemented an experimental protocol for calling external tools from the tactic language.
Matthieu Sozeau developed Russell, an experimental language to specify the behavior of programs with subtypes.
A mechanism to automatically use some specific tactic to solve unresolved implicit has been implemented by Hugo Herbelin.
Laurent Théry’s contribution on strings and Pierre Letouzey and Jean-Christophe Filliâtre’s contribution on finite maps have been integrated to the Coq standard library. Pierre Letouzey developed a library about finite sets “à la Objective Caml”. With Jean-Marc Notin, he extended the library on lists. Pierre Letouzey’s contribution on rational numbers has been integrated and extended.
Pierre Corbineau extended his tactic for solving first-order statements. He wrote a reflection-based intuitionistic tautology solver.
Pierre Courtieu, Julien Forest and Yves Bertot added extra support to reason on the inductive structure of recursively defined functions.
Jean-Marc Notin significantly contributed to the general maintenance of
the system. He also took care of coqdoc
.
Pierre Castéran contributed to the documentation of (co)inductive types and suggested improvements to the libraries.
Pierre Corbineau implemented a declarative mathematical proof language, usable in combination with the tactic-based style of proof.
Finally, many users suggested improvements of the system through the Coq-Club mailing list and bug-tracker systems, especially user groups from INRIA Rocquencourt, Radboud University, University of Pennsylvania and Yale University.
Details of changes in 8.1beta¶
Logic
Added sort-polymorphism on inductive families
Allowance for recursively non-uniform parameters in inductive types
Syntax
No more support for version 7 syntax and for translation to version 8 syntax.
In fixpoints, the { struct ... } annotation is not mandatory any more when only one of the arguments has an inductive type
Added disjunctive patterns in match-with patterns
Support for primitive interpretation of string literals
Extended support for Unicode ranges
Commands
Added "Print Ltac qualid" to print a user defined tactic.
Added "Print Rewrite HintDb" to print the content of a DB used by autorewrite.
Added "Print Canonical Projections".
Added "Example" as synonym of "Definition".
Added "Proposition" and "Corollary" as extra synonyms of "Lemma".
New command "Whelp" to send requests to the Helm database of proofs formalized in the Calculus of Inductive Constructions.
Command "functional induction" has been re-implemented from the new "Function" command.
Ltac and tactic syntactic extensions
New primitive "external" for communication with tool external to Coq
New semantics for "match t with": if a clause returns a tactic, it is now applied to the current goal. If it fails, the next clause or next matching subterm is tried (i.e. it behaves as "match goal with" does). The keyword "lazymatch" can be used to delay the evaluation of tactics occurring in matching clauses.
Hint base names can be parametric in auto and trivial.
Occurrence values can be parametric in unfold, pattern, etc.
Added entry constr_may_eval for tactic extensions.
Low-priority term printer made available in ML-written tactic extensions.
"Tactic Notation" extended to allow notations of tacticals.
Tactics
New implementation and generalization of
setoid_*
(setoid_rewrite
,setoid_symmetry
,setoid_transitivity
,setoid_reflexivity
andautorewite
). New syntax for declaring relations and morphisms (old syntax still working with minor modifications, but deprecated).New implementation (still experimental) of the ring tactic with a built-in notion of coefficients and a better usage of setoids.
New conversion tactic "vm_compute": evaluates the goal (or an hypothesis) with a call-by-value strategy, using the compiled version of terms.
When rewriting H where H is not directly a Coq equality, search first H for a registered setoid equality before starting to reduce in H. This is unlikely to break any script. Should this happen nonetheless, one can insert manually some "unfold ... in H" before rewriting.
Fixed various bugs about (setoid) rewrite ... in ... (in particular bug #5941)
"rewrite ... in" now accepts a clause as place where to rewrite instead of just a simple hypothesis name. For instance:
rewrite H in H1,H2 |- *
meansrewrite H in H1; rewrite H in H2; rewrite H
rewrite H in * |-
will do tryrewrite H in Hi
for all hypothesis Hi <> H.Added "dependent rewrite term" and "dependent rewrite term in hyp".
Added "autorewrite with ... in hyp [using ...]".
Tactic "replace" now accepts a "by" tactic clause.
Added "clear - id" to clear all hypotheses except the ones depending in id.
The argument of Declare Left Step and Declare Right Step is now a term (it used to be a reference).
Omega now handles arbitrary precision integers.
Several bug fixes in Reflexive Omega (romega).
Idtac can now be left implicit in a [...|...] construct: for instance, [ foo | | bar ] stands for [ foo | idtac | bar ].
Fixed a "fold" bug (noncritical but possible source of incompatibilities).
Added classical_left and classical_right which transforms
|- A \/ B
into~B |- A
and~A |- B
respectively.Added command "Declare Implicit Tactic" to set up a default tactic to be used to solve unresolved subterms of term arguments of tactics.
Better support for coercions to Sortclass in tactics expecting type arguments.
Tactic "assert" now accepts "as" intro patterns and "by" tactic clauses.
New tactic "pose proof" that generalizes "assert (id:=p)" with intro patterns.
New introduction pattern "?" for letting Coq choose a name.
Introduction patterns now support side hypotheses (e.g. intros [|] on "(nat -> nat) -> nat" works).
New introduction patterns "->" and "<-" for immediate rewriting of introduced hypotheses.
Introduction patterns coming after nontrivial introduction patterns now force full introduction of the first pattern (e.g.
intros [[|] p]
onnat->nat->nat
now behaves likeintros [[|?] p]
)Added "eassumption".
Added option 'using lemmas' to auto, trivial and eauto.
Tactic "congruence" is now complete for its intended scope (ground equalities and inequalities with constructors). Furthermore, it tries to equates goal and hypotheses.
New tactic "rtauto" solves pure propositional logic and gives a reflective version of the available proof.
Numbering of "pattern", "unfold", "simpl", ... occurrences in "match with" made consistent with the printing of the return clause after the term to match in the "match-with" construct (use "Set Printing All" to see hidden occurrences).
Generalization of induction "induction x1...xn using scheme" where scheme is an induction principle with complex predicates (like the ones generated by function induction).
Some small Ltac tactics has been added to the standard library (file Tactics.v):
f_equal : instead of using the different f_equalX lemmas
case_eq : a "case" without loss of information. An equality stating the current situation is generated in every sub-cases.
swap : for a negated goal ~B and a negated hypothesis H:~A, swap H asks you to prove A from hypothesis B
revert : revert H is generalize H; clear H.
Extraction
All type parts should now disappear instead of sometimes producing _ (for instance in Map.empty).
Haskell extraction: types of functions are now printed, better unsafeCoerce mechanism, both for hugs and ghc.
Scheme extraction improved, see http://www.pps.jussieu.fr/~letouzey/scheme.
Many bug fixes.
Modules
Added "Locate Module qualid" to get the full path of a module.
Module/Declare Module syntax made more uniform.
Added syntactic sugar "Declare Module Export/Import" and "Module Export/Import".
Added syntactic sugar "Module M(Export/Import X Y: T)" and "Module Type M(Export/Import X Y: T)" (only for interactive definitions)
Construct "with" generalized to module paths: T with (Definition|Module) M1.M2....Mn.l := l'.
Notations
Option "format" aware of recursive notations.
Added insertion of spaces by default in recursive notations w/o separators.
No more automatic printing box in case of user-provided printing "format".
New notation "exists! x:A, P" for unique existence.
Notations for specific numerals now compatible with generic notations of numerals (e.g. "1" can be used to denote the unit of a group without hiding 1%nat)
Libraries
New library on String and Ascii characters (contributed by L. Thery).
New library FSets+FMaps of finite sets and maps.
New library QArith on rational numbers.
Small extension of Zmin.V, new Zmax.v, new Zminmax.v.
Reworking and extension of the files on classical logic and description principles (possible incompatibilities)
Few other improvements in ZArith potentially exceptionally breaking the compatibility (useless hypothesys of Zgt_square_simpl and Zlt_square_simpl removed; fixed names mentioning letter O instead of digit 0; weaken premises in Z_lt_induction).
Restructuration of Eqdep_dec.v and Eqdep.v: more lemmas in Type.
Znumtheory now contains a gcd function that can compute within Coq.
More lemmas stated on Type in Wf.v, removal of redundant Acc_iter and Acc_iter2.
Change of the internal names of lemmas in OmegaLemmas.
Acc in Wf.v and clos_refl_trans in Relation_Operators.v now rely on the allowance for recursively non-uniform parameters (possible source of incompatibilities: explicit pattern-matching on these types may require to remove the occurrence associated with their recursively non-uniform parameter).
Coq.List.In_dec has been set transparent (this may exceptionally break proof scripts, set it locally opaque for compatibility).
More on permutations of lists in List.v and Permutation.v.
List.v has been much expanded.
New file SetoidList.v now contains results about lists seen with respect to a setoid equality.
Library NArith has been expanded, mostly with results coming from Intmap (for instance a bitwise xor), plus also a bridge between N and Bitvector.
Intmap has been reorganized. In particular its address type "addr" is now N. User contributions known to use Intmap have been adapted accordingly. If you're using this library please contact us. A wrapper FMapIntMap now presents Intmap as a particular implementation of FMaps. New developments are strongly encouraged to use either this wrapper or any other implementations of FMap instead of using directly this obsolete Intmap.
Tools
New semantics for coqtop options ("-batch" expects option "-top dir" for loading vernac file that contains definitions).
Tool coq_makefile now removes custom targets that are file names in "make clean"
New environment variable COQREMOTEBROWSER to set the command invoked to start the remote browser both in Coq and CoqIDE. Standard syntax: "%s" is the placeholder for the URL.
Details of changes in 8.1gamma¶
Syntax
changed parsing precedence of let/in and fun constructions of Ltac: let x := t in e1; e2 is now parsed as let x := t in (e1;e2).
Language and commands
Added sort-polymorphism for definitions in Type (but finally abandoned).
Support for implicit arguments in the types of parameters in (co)fixpoints and (co)inductive declarations.
Improved type inference: use as much of possible general information. before applying irreversible unification heuristics (allow e.g. to infer the predicate in "(exist _ 0 (refl_equal 0) : {n:nat | n=0 })").
Support for Miller-Pfenning's patterns unification in type synthesis (e.g. can infer P such that P x y = phi(x,y)).
Support for "where" clause in cofixpoint definitions.
New option "Set Printing Universes" for making Type levels explicit.
Tactics
Improved implementation of the ring and field tactics. For compatibility reasons, the previous tactics are renamed as legacy ring and legacy field, but should be considered as deprecated.
New declarative mathematical proof language.
Support for argument lists of arbitrary length in Tactic Notation.
rewrite ... in H
now fails ifH
is used either in an hypothesis or in the goal.The semantics of
rewrite ... in *
has been slightly modified (see doc).Support for
as
clause in tactic injection.New forward-reasoning tactic "apply in".
Ltac fresh operator now builds names from a concatenation of its arguments.
New ltac tactic "remember" to abstract over a subterm and keep an equality
Support for Miller-Pfenning's patterns unification in apply/rewrite/... (may lead to few incompatibilities - generally now useless tactic calls).
Bug fixes
Fix for notations involving basic "match" expressions.
Numerous other bugs solved (a few fixes may lead to incompatibilities).
Details of changes in 8.1¶
Bug fixes
Many bugs have been fixed (cf coq-bugs web page)
Tactics
New tactics ring, ring_simplify and new tactic field now able to manage power to a positive integer constant. Tactic ring on Z and R, and field on R manage power (may lead to incompatibilities with V8.1gamma).
Tactic field_simplify now applicable in hypotheses.
New field_simplify_eq for simplifying field equations into ring equations.
Tactics ring, ring_simplify, field, field_simplify and field_simplify_eq all able to apply user-given equations to rewrite monoms on the fly (see documentation).
Libraries
New file ConstructiveEpsilon.v defining an epsilon operator and proving the axiom of choice constructively for a countable domain and a decidable predicate.
Version 8.0¶
Summary of changes¶
Coq version 8 is a major revision of the Coq proof assistant. First, the underlying logic is slightly different. The so-called impredicativity of the sort Set has been dropped. The main reason is that it is inconsistent with the principle of description which is quite a useful principle for formalizing mathematics within classical logic. Moreover, even in an constructive setting, the impredicativity of Set does not add so much in practice and is even subject of criticism from a large part of the intuitionistic mathematician community. Nevertheless, the impredicativity of Set remains optional for users interested in investigating mathematical developments which rely on it.
Secondly, the concrete syntax of terms has been completely revised. The main motivations were
a more uniform, purified style: all constructions are now lowercase, with a functional programming perfume (e.g. abstraction is now written fun), and more directly accessible to the novice (e.g. dependent product is now written forall and allows omission of types). Also, parentheses are no longer mandatory for function application.
extensibility: some standard notations (e.g. “<” and “>”) were incompatible with the previous syntax. Now all standard arithmetic notations (=, +, *, /, <, <=, ... and more) are directly part of the syntax.
Together with the revision of the concrete syntax, a new mechanism of notation scopes permits to reuse the same symbols (typically +, -, *, /, <, <=) in various mathematical theories without any ambiguities for Coq, leading to a largely improved readability of Coq scripts. New commands to easily add new symbols are also provided.
Coming with the new syntax of terms, a slight reform of the tactic language and of the language of commands has been carried out. The purpose here is a better uniformity making the tactics and commands easier to use and to remember.
Thirdly, a restructuring and uniformization of the standard library of Coq has been performed. There is now just one Leibniz equality usable for all the different kinds of Coq objects. Also, the set of real numbers now lies at the same level as the sets of natural and integer numbers. Finally, the names of the standard properties of numbers now follow a standard pattern and the symbolic notations for the standard definitions as well.
The fourth point is the release of CoqIDE, a new graphical gtk2-based interface fully integrated with Coq. Close in style to the Proof General Emacs interface, it is faster and its integration with Coq makes interactive developments more friendly. All mathematical Unicode symbols are usable within CoqIDE.
Finally, the module system of Coq completes the picture of Coq version 8.0. Though released with an experimental status in the previous version 7.4, it should be considered as a salient feature of the new version.
Besides, Coq comes with its load of novelties and improvements: new or improved tactics (including a new tactic for solving first-order statements), new management commands, extended libraries.
Bruno Barras and Hugo Herbelin have been the main contributors of the reflection and the implementation of the new syntax. The smart automatic translator from old to new syntax released with Coq is also their work with contributions by Olivier Desmettre.
Hugo Herbelin is the main designer and implementer of the notion of notation scopes and of the commands for easily adding new notations.
Hugo Herbelin is the main implementer of the restructured standard library.
Pierre Corbineau is the main designer and implementer of the new tactic for solving first-order statements in presence of inductive types. He is also the maintainer of the non-domain specific automation tactics.
Benjamin Monate is the developer of the CoqIDE graphical interface with contributions by Jean-Christophe Filliâtre, Pierre Letouzey, Claude Marché and Bruno Barras.
Claude Marché coordinated the edition of the Reference Manual for Coq V8.0.
Pierre Letouzey and Jacek Chrząszcz respectively maintained the extraction tool and module system of Coq.
Jean-Christophe Filliâtre, Pierre Letouzey, Hugo Herbelin and other contributors from Sophia-Antipolis and Nijmegen participated in extending the library.
Julien Narboux built a NSIS-based automatic Coq installation tool for the Windows platform.
Hugo Herbelin and Christine Paulin coordinated the development which was under the responsibility of Christine Paulin.
Details of changes in 8.0beta old syntax¶
Logic
Set now predicative by default
New option -impredica